Trust Me, I Might be a Medical Device: The Problem with AI Scribes

Trust Me, I Might be a Medical Device: The Problem with AI Scribes | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Trust Me, I Might be a Medical Device: The Problem with AI Scribes Claire Dady, Henry Potts, Fiona Stevenson, Patricia Schartau This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8994700/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background AI-enabled digital medical scribes are increasingly promoted as a solution to documentation burden and burnout in UK primary care. Recent NHS guidance indicates that many such tools may fall within the scope of UK Medical Devices Regulations, yet there is a lack of support for clinicians to verify regulatory status or data governance claims. Objective To identify AI-powered health care digital scribe (ambient voice technology, AVT) products available to UK clinicians and evaluate the transparency of their regulatory, data-protection and NHS-readiness claims. Methods Between 1–15 November 2025 we conducted a structured web-based review using search engines, NHS innovation and procurement portals, regulatory databases (including the MHRA Public Access Registration Database), commercial catalogues and vendor materials. Products were eligible if they were marketed as healthcare AI scribes or consultation assistants that transcribe and/or summarise clinician–patient interactions. For each tool, we extracted information on features, GDPR statements, MHRA registration, claimed NHS use and access pathways. Results We identified 118 AI-enabled health care AVT products marketed as AI scribes. Only nine (7%) explicitly claimed Class I medical device status; of these, six could be definitively located on the MHRA register under the same or a clearly linked legal entity. An explicit GDPR or data-protection statement was present for 88 products (73%), but supporting evidence (e.g. DPIAs, UK-specific safeguards) was rarely available. In terms of market accessibility, 51 tools (42%) required direct vendor contact or a demonstration, 46 (38%) showed clear UK or NHS availability, 15 (12%) were not accessible to UK users, and access status was unclear for nine (7%). Conclusions There is a marked mismatch between emerging NHS expectations for AI-enabled medical scribes and the transparency of current market offerings. Clinicians may struggle to reliably verify which tools are registered medical devices, how patient data are processed, or whether NHS clinical safety standards are met. Although NHS England has since introduced a national self-certified Ambient Voice Technology (AVT) supplier registry to improve visibility of regulatory and governance compliance, this was published after completion of data collection for this study. While the registry represents an important step toward strengthening oversight, its impact will depend on supplier participation, accuracy of self-certification, and effective dissemination to clinicians and healthcare organisations. Medical Informatics AI medical scribes ambient voice technology digital health regulation UK Medical Devices Regulations MHRA registration NHS governance GDPR compliance clinical safety standards software as a medical device healthcare AI transparency Introduction Recent advances in large language models (LLMs) have driven the development of AI-based digital medical scribes, systems designed to capture clinician, patient interactions and generate structured documentation for inclusion in electronic health records. Depending on the vendor, these tools provide real-time transcription, automated summarisation, SNOMED coding, and the generation of formatted consultation notes or referral letters. The NHS Ten-Year Plan (2025) identifies artificial intelligence as a central enabler of digital transformation, explicitly referencing AI-enabled documentation assistants as a route toward a more data-driven and efficient healthcare system. In parallel, NHS England issued guidance in April 2025 confirming that AI scribes with specific functionalities may fall within the scope of the UK Medical Devices Regulations 2002, requiring registration with the Medicines and Healthcare products Regulatory Agency (MHRA) as software medical devices. In January 2026, NHS England introduced an AI-Enabled Ambient Voice Technology (AVT) Registry hoping to improve transparency and enable suppliers to self-certify compliance with NHS standards, including DTAC and MHRA requirements. Persistent weaknesses in oversight across digital health technologies reinforce the need for such mechanisms. Many tools currently deployed within the NHS lack robust evidence of safety, benefit, or regulatory compliance, reflecting fragmented accountability and insufficient post-market evaluation (Oskrochi et al., 2025). Despite these policy advances, the UK market for AI scribes remains uncoordinated. Many products are marketed directly to clinicians or small practices without demonstrable alignment to UK regulatory or clinical safety standards. This study presents a structured review of AI scribe products currently available to UK clinicians. It analyses publicly available vendor information and MHRA registration data to assess the transparency of regulatory claims and the challenges clinicians face in verifying compliance. Speech-to-text transcription tools have long been used in healthcare, but the latest generation of AI scribes marks a significant shift in capability. According to NHS England’s April 2025 guidance, newer tools leveraging LLMs to summarise consultations are no longer considered passive documentation aids. Depending on their functionality, they may fall within the scope of the UK Medical Devices Regulations 2002 (NHS England, 2025b). The Medicines and Healthcare products Regulatory Agency had previously, on 3 February 2025, concluded that scribes in mental health likely fall within this remit. The NHS 10-Year Health Plan reinforces this evolution, positioning ambient AI (AVT) and digital assistants as central to the NHS’s digital-transformation strategy. Pilot evaluations cited in the plan report that AI documentation tools have reduced administrative workload by over 50%, enabling staff to focus more on patient care and improving service throughput (NHS England, 2025a). AI tools are increasingly present in UK primary care, with around one-third of GPs already using them, yet meaningful impact is not assured. Adoption remains uneven, shaped by differing local policies in which some ICBs encourage cautious piloting while others promote more proactive use of approved systems (Kumpunen et al., 2025). However, these innovations raise complex regulatory and ethical challenges. The MHRA defines software as a medical device as any software intended for a medical purpose, including diagnosis, treatment, or prognosis. Even if a developer does not explicitly claim medical use, a tool’s interface, marketing, or deployment context can bring it within regulatory scope. Features such as automated summarisation, decision support, or clinical coding particularly when not easily verifiable by clinicians, increase the likelihood of classification as a medical device (MHRA, v1.10f). Oskrochi et al. (2025) identified persistent deficiencies in compliance, governance, and post-market monitoring across a wide range of digital health tools currently in use. Their analysis demonstrated that many technologies are implemented without clear evidence of safety, benefit, or regulatory alignment, and that responsibility for compliance remains fragmented across multiple agencies. The authors concluded that NHS oversight mechanisms have not kept pace with the rapid evolution of digital innovation, resulting in gaps in accountability and transparency. In response to these challenges, NHS England has now introduced a self-certified AVT supplier registry, intended to act as a national reference point for organisations and clinicians seeking assurance on regulatory and governance alignment https://transform.england.nhs.uk/digitise-connect-transform/digitising-the-frontline/ambient-voice-technology-self-certified-supplier-registry/ . The registry enables suppliers to declare compliance with NHS requirements, including DTAC, clinical safety standards, and (where applicable) MHRA medical device registration. While this represents an important structural intervention, the registry relies on self-certification rather than independent verification, and its effectiveness will depend on supplier participation, accuracy of declarations, and visibility to end-users. At present, there is limited evidence regarding clinician awareness of the registry or its integration into routine procurement, governance, or audit processes. Regulatory classification carries significant implications. Products meeting the definition of a medical device must be registered with the MHRA, comply with UK Conformity Assessed (UKCA) or CE-marking requirements, and adhere to NHS clinical-safety standards (DCB0129 and DCB0160). These obligations may present barriers for smaller developers or overseas vendors unfamiliar with UK-specific regulation. The 10-Year Plan acknowledges these challenges, committing to a “faster, predictable, and risk-proportionate” regulatory framework for AI and software-as-a-medical-device, developed jointly by the MHRA and the Department for Science, Innovation and Technology (NHS England, 2025a). Although the statutory MHRA registration fee for software devices is modest, currently £261 (rising to £300 annually in 2026), the true cost of achieving full regulatory compliance is substantial. When accounting for engineering design, risk management, regulatory documentation, clinical validation, and conformity assessment, the average development cost for a Class II medical device (including digital and software-only systems) ranges between £1.6 million and £4 million (MedTech Europe, 2024), albeit software-only devices typically occupy the lower end of this range. There is a high turnover of small providers in the emerging digital-scribe (AVT) market. The advent of accessible LLM frameworks and open developer APIs has lowered the barrier to entry, enabling start-ups and independent developers to produce AI documentation tools with limited resources. Many of these vendors market directly to clinicians and small practices via web platforms, promoting their tools as workflow aids rather than regulated medical devices. However, the cost and complexity of full Class II certification, including quality-management systems and clinical-performance evidence, may be prohibitive for all but the largest or most established suppliers. Professional responsibility nevertheless remains with the clinician, who must review and verify AI-generated outputs before they are entered into the patient record. Failure to identify inaccuracies could expose clinicians to liability if harm occurs, even when errors originate from the AI system (Nicholson et al., 2024). Automation bias where clinicians over-rely on plausible AI outputs has been recognised as an emerging safety concern (Abdelwanis et al., 2024). Further governance questions surround patient consent and data transparency. While some providers verbally inform patients when AI scribes are used, consent models vary. Under UK GDPR and the Data Protection Act 2018, patients have the right to understand how their data are processed, and professional guidance from the GMC and National Data Guardian emphasises the need for transparency. Although the NHS envisions a unified, digital-first health record, it is not yet standard practice to mark AI-generated documentation within patient records. This omission has implications for accountability, auditability, and trust (Mennella et al., 2024). Methods This study aimed to identify AI-powered digital medical scribes available to clinicians in the United Kingdom and assess the transparency of their regulatory status. The analysis focused on tools that are marketed as using AI to transcribe or summarise patient consultations and are promoted for use in healthcare settings. Eligibility Criteria To be included in the review, a product had to: Be marketed as an AI-powered digital scribe or consultation assistant in a health care setting including mental health/therapy; Offer transcription, summarisation, or clinical documentation support based on patient–clinician interaction; Tools offering only basic dictation or voice recognition (e.g. standalone speech-to-text apps) were excluded unless they also included generative summarisation or structured documentation functionality. Search Strategy The search was conducted between 1–1 5 Nov 2025 using a combination of formal and informal sources to identify products available to UK clinicians. This included: Structured web searches (Google) using terms such as “AI scribe UK,” “digital medical scribe NHS,” “ambient documentation tool,” and “LLM scribe healthcare”; NHS procurement websites, innovation portals, and accelerator programme listings; Publicly accessible regulatory databases including the MHRA Public Access Medical Device Registration System; U.S healthcare technology data base - Elion; Company websites, marketing materials, press releases, and white papers; Professional discussion forums, including Reddit threads and X posts; LinkedIn company pages, employee posts, and industry analyst commentary. This approach generated a list of 121 AI scribe tools potentially available to UK clinicians. These included both UK-developed and international tools, some of which are in active NHS deployment or pilot phases. Data Extraction and Review For each tool, data was manually extracted from public sources, including: Product features (e.g. transcription, summarisation, coding); Regulatory status (e.g. MHRA registration); Integration claims with UK EHR systems and NHS compliance (e.g. EMIS, SystmOne); Data handling practices (e.g. patient data storage, audio retention, anonymisation); Patient consent processes and governance indicators, if any (e.g. documentation of AI use, consent procedures, audit trails). Where regulatory status was unclear or unverifiable via official databases, this was recorded as a key finding. Limitations This review has several important limitations. First, data collection was completed prior to the publication of the NHS England Ambient Voice Technology (AVT) self-certified supplier registry in January 2026. As a result, the analysis does not account for supplier participation in, or compliance declarations made through, this registry. The findings therefore represent a snapshot of the market immediately before the introduction of a national assurance mechanism and may over-estimate the current degree of opacity in regulatory signalling. The MHRA registration database is not searchable by product function, AI capability, or clinical application, which makes it difficult to confirm whether a product is in scope or appropriately registered. Some tools may be undergoing registration, or may self-certify as Class I devices under UK Medical Devices Regulations 2002, which does not require independent third-party assessment. As stated in MHRA guidance: “Registration of medical devices with the MHRA... does not represent any form of accreditation, certification, approval or endorsement by the MHRA. Therefore, it is not permitted to make any claims to this effect…” (MHRA, 2023). As such, the presence of a tool on the register should not be interpreted as confirmation of compliance with NHS safety, performance, or clinical governance expectations. Additionally, several vendors do not publish regulatory documentation or safety information on their websites, further limiting the ability to verify claims. Results Overview A total of 118 AI-enabled ambient scribing (AVT) products were identified as being marketed to, or accessible by, UK clinicians. These include UK-based and international suppliers offering capabilities such as ambient voice capture, real-time transcription, generative summarisation, structured coding, and automated referral generation. Claims of regulatory compliance, NHS readiness, and data-handling transparency varied widely. Scribe Name MHRA Device Registration UK Access Availability Compliance Notes Abridge No Contact required Privacy policy legal, no direct GDPR citation Accel eq No No Some Accel privacy docs claim GDPR, check clinical context Aesthetics 360 No Contact required UK sector guidance, no direct product claim aide health Yes (Class 1) Contact required NHS DCB0129/0160 clinical safety standards, GDPR data protection. Aiva No Contact required GDPR, privacy statement, EU/UK data rights Ambience Scribe No Contact required NHS-required GDPR, needs direct confirmation Andy No Contact required GDPR, HIPAA, SOC2, security audits Anima/ Annie (UK) Yes (class 1 claimed), could not find on pard Yes UK GDPR,Use in nhs reported, DPA, DTAC, DSPT, DCB0129 Augmedix (USA) No No HIPAA, de-identification, US focus Augnito (UK/IND) Yes (class 1) (ScribeTech) NHS dictation UK GDPR, NHS, local hosting, security, DPA Aura AI Scribe No Yes GDPR, HIPAA, SOC2, consent AutoMynd No Contact required GDPR, HIPAA, SOC2, DPO, consent, EU hosting Avo No Contact required GDPR, HIPAA, CCPA, anonymization, audit Avodahmed No Contact required GDPR, patient consent, subject rights, industry audits BastionGPT No Yes GDPR, HIPAA, SOC2, UK/EU legal review Bells ai No No GDPR, HIPAA, SOC2, PIPEDA, UK data controls Blueprint No Yes References OpenAI's privacy; not product-level GDPR Care cortex No Contact required GDPR, HIPAA, ISO, DPO, subject rights CarePilot No Contact required GDPR, HIPAA, UK DPA, Aus Privacy Act, DPO Chartnote No Contact required GDPR, HIPAA, PIPEDA, SOC2, staff training, logging CircleHealth AI No Contact required HIPAA, Joint Commission, lacks GDPR Clinicminds No Contact required GDPR, HIPAA, PIPEDA, privacy, consent management Commure Ambient AI No Contact required GDPR, HIPAA, SOC2, end-to-end encryption Continuous care No Contact required GDPR, HIPAA, EU/UK hosting, right to be forgotten Conveyor AI No Contact required GDPR, HIPAA, SOC2, DPA/EU/UK data addendum Corti No Yes GDPR, HIPAA, ISO27001, SOC2, medical device certs DeepCura AI (USA/CAN) No Contact required GDPR, HIPAA, SOC2, DPO, ongoing audits DeepScribe (USA) No Unclear HIPAA, SOC2 only Deliberate ai No Contact required No public policy, GDPR not found Diktamen no Yes DTAC, DCB0129, DSPT, Cyber esentials, ISO 27001, ISO 9001, ISO14001 Dragon Medical One (USA) No NHS use claimed Azure UK/EU hosting, ISO27001, DCB0129, NHS compliance Eleos No Contact required UK privacy, GDPR/UK GDPR, subject rights, DPO, audits Empathia ai No Yes GDPR, HIPAA, ISO27001, privacy basis Enzohealth No Contact required GDPR, HIPAA, ISO27001, SOC2 Expert Ai No No GDPR for Cogito, DPO support, EU bases only for that prod. Ezy scribe No No No public GDPR documentation Fireflies ai No No GDPR, SOC2, HIPAA, US hosting, enterprise EU option Freed AI (USA) No Yes GDPR, HIPAA, HITECH, audits, US/EU privacy Glass health No Contact required HIPAA, US-centric only Halo scribe No Contact required No public GDPR evidence Health orbit Yes Yes, NHS use claimed GDPR, UK/EU data handling, privacy framework Heidi Health (AUS) Yes (Class 1) Yes, NHS use claimed GDPR, HIPAA, NHS, ISO27001, UK hosting Innovaccer InScribe No Yes GDPR, HIPAA, SOC2, Trust Center, EU/UK Insight aura AI scribe No Yes GDPR, HIPAA, SOC2, ISO27001, UK Cyber Essentials Iscribe No No (not supported) HIPAA-only focus, no GDPR evidence Jessica by Combine Health No No HIPAA, CMS, lacking explicit GDPR Jotpsych No Yes GDPR, HIPAA, SOC2 Type 2, audit, deletion Kiwipen (UK) No Yes GDPR, ICO, Cyber Essentials, DSPT, DCB0129, DTAC Knowtex Yes Yes HIPPA, SOC 2, GDPR, iso 27001 Lexacom Scribe Yes States class one on website unable to find on pard Contact required GDPR, HIPAA, ISO, UK DPA, DPO, EU/UK data controls Lucus ai No Yes GDPR, HIPAA, SOC2, UK data residency Lyrebird Health (AUS) Yes (Class 1) Yes, NHS use claimed GDPR, NHS DTAC, DCB0129, HIPAA, Australia Privacy Act Mariana ai No No HIPAA only, not GDPR for healthcare Marvix No Contact required GDPR, SOC2, EEA data, subject rights, breach notification Mdhub No Yes GDPR, HIPAA, SOC2, security audits Mediary No Contact required No privacy or public GDPR evidence, not AVT scribe MEDILIT No Yes GDPR, HIPAA, SOC2, product claim MediSync No Contact required GDPR, HIPAA, Texas regional, Elion Health confirmed Mentalyc No No GDPR, HIPAA, PHIPA, PIPEDA, rights, audit Mircosoft DAX Yes website states class 1 could not find on pard Contact required GDPR, HIPAA, HITRUST, SOC2, SCCs, EU/UK hosting Mutuo AutoScribe (CAN) No Yes GDPR, HIPAA, SOC2, PHIPA, PIPEDA, UK/EU support Nabla Copilot (FRA/USA) No Yes GDPR, HIPAA, SOC2, ISO27001, EU/US hosting NextGen Ambient Assist No Contact required GDPR, HIPAA, SOC2, Elion Health review Nextvisit AI No No Not GDPR compliant; US only, not for UK/EU (privacy stmt) Nora note No Yes GDPR, privacy, consent, UK/EU legal Nudge No Yes GDPR, minimization, data rights, consent OmniMD AI Medical Scribe No Contact required GDPR, HIPAA, CCPA, SOC2, audits, EU/UK controls OneChart No Yes No public GDPR evidence or policy OneChart (CAN) No Yes No public GDPR evidence or policy OneLine Health no Contact required GDPR, HIPAA, PIPEDA, NHS, privacy/erasure Opus No Contact required GDPR, HIPAA, SOC2, product privacy solutions Oracle Clinical Digital Asst. No Contact required GDPR, HIPAA, SOC2, DPA, SCC, EU privacy safeguards OrbDoc No Contact required GDPR, HIPAA, SOC2, ISO27001, NHS DSPT, UK/EU Pabau - Echo AI No Yes GDPR, HIPAA, consent/audit features PatientNotes No Yes GDPR, HIPAA, UK DPA, Aus Privacy Act, audit logs Playback health No Yes GDPR, HIPAA, SOC2, explicit EU controls Plena Health AI Scribe No Contact required GDPR, HIPAA, SOC2, UK/EU privacy controls Giiq No No GDPR, staff training, privacy, subject rights Regard No Contact required HIPAA, SOC2 Reteta No Contact required No public GDPR privacy policy available Revmax No Contact required GDPR, legal processing EU/UK, DPA, data rights River Records No Contact required GDPR, UK DPA, UK/EU data, subject rights S10.AI No Contact required GDPR, HIPAA, ISO, PIPEDA, EU hosting, DPA available Sayvant No Contact required GDPR, HIPAA, SOC2, audits, EU/UK subject rights Scribble No Error 403 forbidden message Privacy error, possibly inaccessible in UK Scribe america speke No No HIPAA, US-centric only Scribeberry No Contact required GDPR, HIPAA, SOC2, ISO, audits, Canadian compliance scribematic No Yes GDPR, HIPAA, ISO, CCPA, NHS guides, UK/EU privacy ScribeMD No Contact required GDPR, HIPAA, SOC2, EU/UK subject rights ScribePT (USA) No Contact required US privacy only ScribeRyte No Contact required GDPR, HIPAA, SOC2, ISO27001, privacy controls Shine AI beam health No Contact required GDPR, HIPAA, SOC2, independent audits, privacy policies SimboAlphus No No Some Symbolab legal basis, not healthcare certain SKRIBER No Yes Data protection, legal compliance Soap suds No Yes No GDPR for health context SOAPNoteAI No Yes GDPR, HIPAA, SOC2, Cyber Essentials, ISO, PIPEDA Solventum fluency direct no Contact required GDPR, HIPAA, SOC2, TX-RAMP, Cyber Essentials, ISO 27001 Speak matic No Unclear No explicit GDPR found Splose no Yes GDPR, UK GDPR, ISO 27001, Health data encryption, SOC2, SSO/2FA Sporo AI Scribe No Yes GDPR, HIPAA, SOC2, UK Cyber Essentials, NHS DSPT, ISO Suki AI (USA) No No GDPR, HIPAA, anonymization, security, user consent Sully u.s No Unclear GDPR, ISO 27001, SOC2, EEA/UK compliance Supanote AI No Yes GDPR, HIPAA, PHIPA, PIPEDA, consent, audits SwiftyNote no Unclear. Code never sent for log in after request GDPR, HIPAA, AWS, privacy/security T-Pro No Yes NHS DCB0129/0160, DTAC, DSPT Tali AI (CAN) No Contact required HIPAA, SOC2, PHIPA/PIPA, lacks explicit GDPR Tandem Yes (Class1) Yes NHS use UK GDPR, NHSE, DPA, NHS toolkit, DPIA, CyberEssentials TheraPulse No Yes GDPR, HIPAA, SOC2, Elion Health confirmed Tortus (UK) Yes (Class 1) Yes NHS use GDPR, CyberEssentials Plus, DTAC, DSPT, DCB0129, NHS Triple time No Yes No GDPR evidence found in public docs Twofold Health (USA) No Yes HIPAA, no public GDPR evidence Upheal No Yes GDPR, HIPAA, SOC2, PHIPA, PIPEDA, UK/EU hosting Veradigm Ambient Scribe No Contact required GDPR, HIPAA, SOC2, EU/UK compliance Vero No Contact required GDPR, ISO27001, SOC2 Type II, HIPAA, external audits Videra Health No Contact required GDPR, HIPAA, SOC2, privacy policy EU/UK Wavo Health (CAN) No Yes GDPR, HIPAA, SOC2, no audio storage, UK hosting Zanda No Yes GDPR, HIPAA, ISO, UK DPA, DPO, EU/UK data controls Results Regulatory verification (MHRA) Manual verification against the Medicines and Healthcare products Regulatory Agency (MHRA) Public Access Registration Database (PARD) https://pard.mhra.gov.uk/ identified nine of 118 AI-enabled AVT products (8%) that claimed Class I medical device status. Of these nine products, six (5% of the identified AVT products) could be definitively located on PARD under the same or a clearly linked manufacturer or legal entity name at the time of review. For the remaining three products (3% of all AVT products), no corresponding PARD entry could be identified under the marketed product name. While these products may be registered under alternate legal entities or parent companies, this could not be confirmed using publicly available records. Table 1 Publicly stated and verifiable MHRA regulatory status of AI-enabled ambient voice technology products marketed to UK clinicians (N = 118) Regulatory status Count % of total Claims MHRA Class I medical device status 9 8 Verified on MHRA PARD under same or clearly linked legal entity 6 5 Claimed Class I status but not verifiable on PARD under marketed name 3 3 No claim of MHRA Class I status 109 92 Verification was defined as a clear and defensible link between the marketed product and a registered legal entity on the MHRA Public Access Registration Database (PARD). Products not locatable under the marketed name may be registered under alternative legal or trading names. Data protection and GDPR compliance Of the 118 AVT products reviewed, 88 (75%) provided an explicit statement of compliance with the General Data Protection Regulation (GDPR) or UK data protection legislation on their websites or in associated documentation. Only a minority of these products linked supporting evidence, such as Data Protection Impact Assessments or references to specific UK GDPR provisions. The remaining products either provided general statements relating to privacy or security without explicit reference to GDPR or did not provide sufficient publicly available information to determine compliance. Table 2 Publicly stated GDPR and data protection compliance among AI-enabled ambient voice technology products (N = 118) Data protection status Count % of total Explicit GDPR or UK data protection compliance statement 88 75 Generic privacy or security statements without explicit GDPR reference 23 19 No clear data protection information identified 7 6 Explicit compliance was defined as a direct reference to GDPR or UK data protection legislation in publicly available product documentation or websites. UK access and deployment Most AVT products demonstrated some degree of potential accessibility to the UK market, although access pathways varied considerably. Forty-nine products (42%) required direct vendor contact or a product demonstration prior to access, indicating possible UK availability but limited public transparency. Forty-three products (36%) were clearly accessible to UK users or reported active or planned deployment within UK or NHS settings. Fifteen products (13%) explicitly stated that they were not available to UK users or were not accesessible from a UK IP address, while 11 products (9%) had unclear or indeterminate access status due to insufficient or inconclusive publicly available information. Table 3 UK access and deployment characteristics of AI-enabled ambient voice technology products (N = 118) Access category Count % of total Interpretation Contact or demonstration required 49 42 Potential UK availability but limited public transparency Accessible to UK users or NHS 43 36 Clear UK access or evidence of NHS use Not accessible to UK users 15 13 Explicit geographical or market restriction Unclear or indeterminate 11 9 Insufficient or inconclusive public information A substantial proportion of AVT products referenced current or intended use within the NHS, commonly in the context of pilot projects, evaluations, or integrations with established clinical systems such as EMIS or SystmOne. Notably, all products that claimed MHRA Class I medical device status also reported current or planned use within NHS settings. Several products without verifiable Class I registration displayed NHS logos or referenced NHS partnerships on their websites. While such references suggest market intent, they were not consistently supported by verifiable evidence of formal contracts or regulatory approval. In addition, a number of products appeared to be primarily aligned with international regulatory frameworks, particularly those applicable in the United States, rather than explicitly addressing UK-specific regulatory and governance requirements. Observed Trends and Common Gaps Brand–manufacturer discrepancies in MHRA listings can limit traceability and slow due diligence checks. In addition, cross-border data processing arrangements are often not clearly specified on supplier websites, including whether appropriate UK safeguards are in place when data are stored or processed outside the UK. Marketing assurances such as claims of being “secure” or “HIPAA compliant” are frequently presented without sufficient detail to meet NHS assurance requirements. These results reinforce the case for a central NHS mechanism such as the AVT Registry to provide verifiable evidence, reduce duplication of effort across ICBs, and support consistent, safe procurement decisions. Discussion This review highlights a significant gap between the regulatory expectations outlined in NHS and MHRA guidance and the transparency of AI scribe tools currently available to UK clinicians. Despite the 2025 NHS statement that high-functionality AI scribes may fall under the UK Medical Devices Regulations 2002, only a minority of tools reviewed had verifiable MHRA registration, UKCA/CE certification, or published clinical safety documentation. For frontline clinicians, this creates uncertainty and potential professional risk. AI scribes that generate summaries, add clinical codes, or format notes for upload into patient records are not merely passive transcription tools. If a tool meets the legal definition of a medical device, clinicians using unregistered versions may bear liability in the event of harm, particularly if due diligence around regulation and safety was not exercised. Moreover, automation bias may lead clinicians to over-rely on AI-generated notes, especially under time pressure or when the output appears plausible. Yet the appeal of AI scribes is undeniable. These tools offer benefits to clinicians and patients alike, particularly for healthcare professionals with high administrative burdens. In the context of increasingly time-constrained consultation models, tools that can reduce cognitive load and streamline note-taking hold real promise for improving care delivery (Tierney et al., 2025). The challenge lies in the pace of development. Rapid advances in generative AI, including the integration of large language models such as GPT-4, have outstripped the capacity of existing regulatory systems to adapt. Today’s AI scribes go far beyond speech-to-text, generating structured summaries, coded entries, and even referral letters within seconds (Sollof, 2025). For clinicians, this raises the question of how they can critically evaluate and safely deploy tools whose regulatory status, audit mechanisms, and error-handling processes remain unclear. This concern becomes especially pressing when AI scribes are embedded directly into electronic patient record (EPR) systems. Errors in transcription, summarisation, or clinical coding may be silently propagated across systems, leading to downstream risks. As more patients gain access to their health records, inaccuracies could generate confusion, complaints, or requests for correction. Misattributed diagnoses or incorrect codes could directly contribute to patient harm and may also create new burdens for data protection officers as well as compromising clinician–patient trust. These findings also suggest that while the majority of AVT scribes have some pathway to the UK market, accessibility and verifiable compliance remain inconsistent. Nearly half of all identified tools require vendor engagement in the form of an online demonstration and discussion with a sales agent before access is granted. Fewer than ten could demonstrate confirmed medical device registration. This suggests a market still in an early stage of regulatory maturity, where international vendors are open to UK adoption but have not fully aligned with UK-specific safety, data governance, or registration standards this does not stop them trying to sell into the UK market. Vendor engagement may, however, serve as a modest safeguard. The need to arrange a demonstration or contact a sales team introduces a degree of friction that could prevent clinicians from adopting tools without adequate due diligence or confirmation of GDPR compliance. In that sense, the “demo/contact required” model may help mitigate inappropriate or premature use. Nevertheless, discussions with several US-based vendors indicated a strong interest in selling to the UK market despite limited understanding of UK regulatory expectations. This lack of awareness could present risks for clinicians who assume these products meet local compliance standards when, in fact, they may only adhere to US frameworks such as HIPAA (Health Insurance Portability and Accountability Act). In parallel, the NHS would benefit from a national audit framework for AI-generated documentation. Standardised approaches are needed to determine how frequently outputs should be reviewed, what constitutes an acceptable error threshold, and how the role of clinical safety officers should evolve. Without such benchmarks, quality assurance will remain fragmented and reactive. The introduction of the NHS England AVT self-certified registry represents a potentially important inflection point in this landscape. If widely adopted and appropriately maintained, the registry could reduce duplication of due diligence across integrated care boards, support more consistent procurement decisions, and provide clinicians with a clearer route to verifying baseline compliance. However, for this potential to be realised, the registry must be visible, trusted, and actively embedded within NHS digital governance workflows. Without proactive communication, integration into existing NHS systems, or linkage to professional guidance, there is a risk that the registry remains peripheral, particularly as commercial vendors continue to market directly to clinicians outside formal procurement pathways. Finally, while initiatives such as RADIANT (Regulatory and Innovation Network for AI Technologies) and CERSI AI (Centre of Excellence for Regulatory Science & Innovation in AI) are emerging to address the governance gap (Bray, 2025), their impact will depend on strong inter-agency coordination, visible national leadership and funding. A transparent, central register of compliant tools, such as the planned NHS AVT Registry, would provide clinicians and organisations with a trusted reference point. In the absence of such infrastructure, responsibility continues to fall on individual users to assess regulatory risk, a task that remains difficult in the current environment. Conclusion AI digital scribes offer considerable promise in reducing documentation burden, improving clinical workflows, and enhancing the quality of records. However, this review has shown that despite clear NHS guidance, the current regulatory landscape lacks transparency and consistency. Many tools marketed to UK clinicians do not clearly disclose their regulatory status. The recent introduction of a national NHS self-certified AVT supplier registry is a welcome and necessary development, signalling recognition of the governance gap identified in this study. However, its late arrival relative to market growth, reliance on self-certification, and limited visibility underscore the need for continued scrutiny. To support safe, accountable adoption of AI scribes, registry participation must be coupled with clearer guidance for clinicians, improved discoverability, and alignment with audit and clinical safety processes. Without these measures, responsibility for managing regulatory risk will continue to fall on individual users, undermining both patient safety and clinician confidence in emerging AI documentation technologies. Without such action, the NHS risks entering a period of unchecked adoption, where the benefits of AI scribes may be undermined by avoidable risks to patients, clinicians, and healthcare systems alike. Declarations Conflicts of Interest/Funding HWWP is partly funded as part of the RADIANT CERSI by the Medical Research Council, Innovate UK, Medicines and Healthcare products Regulatory Agency, and Office for Life Sciences. Views presented are the authors and do not represent those of the funding organisations. HWWP provides or has provided consultancy on digital health evaluation for Flo Health Inc. and Thrive Therapeutic Software Ltd. He has PhD students in the field employed by or previously employed by, and with fees paid by or previously paid by AstraZeneca, Patients Know Best, and BetterPoints Ltd. References Abdelwanis, M., Alarafati, H.K., Tammam, M.M.S. and Simsekler, M.C.E. (2024). Exploring the risks of automation bias in healthcare artificial intelligence applications: A Bowtie analysis. Journal of Safety Science and Resilience , [online] 5(4), pp.460–469. doi:https://doi.org/10.1016/j.jnlssr.2024.06.001. Abridge.com. (2025). Generative AI for Clinical Conversations | Abridge . [online] Available at: https://www.abridge.com/ [Accessed 15 Nov. 2025]. Accurx.com. (2022). Accurx Scribe: Privacy and Security FAQs | Accurx Help Centre . [online] Available at: https://support.accurx.com/en/articles/11156712-accurx-scribe-privacy-and-security-faqs [Accessed 15 Nov. 2025]. Aesthetics360.com. (2025). aesthetics360 . [online] Available at: https://aesthetics360.com/ [Accessed 16 Nov. 2025]. AI Medical Scribe for Clinicians | Wavo Health. (2025). AI Medical Scribe for Clinicians | Wavo Health . [online] Available at: https://wavo.health/en [Accessed 15 Nov. 2025]. AI Medical Scribe for Healthcare & Medicine | Skriber. (2025). Home - AI Medical Scribe for Healthcare & Medicine | Skriber . [online] Available at: https://skriber.com/ [Accessed 15 Nov. 2025]. Aide.health. (2018). Aide for Clinicians — Rebuilding chronic disease care with AI . [online] Available at: https://aide.health/products/aide [Accessed 15 Nov. 2025]. Aiscribee.com. (2025). Home | AI Medical Scribe & EMR Assistant | ScriBee . [online] Available at: https://aiscribee.com/ [Accessed 15 Nov. 2025]. Ali, M., Fantesini, A., Morcella, M.T., Ibba, S., Gennaro D'Anna, Fazzini, D. and Papa, S. (2023). Adoption of AI in Oncological Imaging: Ethical, Regulatory, and Medical-Legal Challenges. Critical Reviews TM in Oncogenesis , [online] 29(2), pp.29–35. doi:https://doi.org/10.1615/critrevoncog.2023050584. Ambiencehealthcare.com. (2025). Ambience Healthcare . [online] Available at: https://www.ambiencehealthcare.com/ [Accessed 15 Nov. 2025]. Anaiva.io. (2025). AI Medical Scribe Software | Automated Clinical Documentation | Aiva . [online] Available at: https://www.anaiva.io/ [Accessed 15 Nov. 2025]. Animahealth.com. (2025). Anima | Anima - Full Compliance with NHS England AVT Requirements . [online] Available at: https://www.animahealth.com/blog/anima-full-compliance-with-nhs-england-avt-requirements [Accessed 15 Nov. 2025]. AppliedAI - Opus. (2025). Opus - AI Workflow and Business Automation Platform | Opus . [online] Available at: https://www.opus.com/ [Accessed 15 Nov. 2025]. Augmedix.com. (2025). Terms of Agreement | Augmedix . [online] Available at: https://www.augmedix.com/terms-of-agreement/current [Accessed 15 Nov. 2025]. Augnito.ai. (2025). Augnito | Enhance Clinical Documentation with Voice AI . [online] Available at: https://augnito.ai/ [Accessed 15 Nov. 2025]. Automynd.com. (2024). Orchestrating AI for Home-Based Care Operations . [online] Available at: https://www.automynd.com/ [Accessed 15 Nov. 2025]. Avaamo. (2021). Avaamo . [online] Available at: https://avaamo.ai/ [Accessed 15 Nov. 2025]. Avomd.com. (2024). AI Software for Healthcare | Avo . [online] Available at: https://www.avomd.com/ [Accessed 15 Nov. 2025]. Bastiongpt.com. (2025). ChatGPT for Healthcare | Medical GPT with HIPAA Compliance . [online] Available at: https://bastiongpt.com/ [Accessed 15 Nov. 2025]. Beam Health. (2025). Beam Health | End-to-End EMR Automation for Your Practice. [online] Available at: https://www.beam.health/ [Accessed 15 Nov. 2025]. Beer, H. (2025). CERSI-AI: shaping safe and effective AI in healthcare . [online] Hospital Healthcare Europe. Available at: https://hospitalhealthcare.com/in-depth/health-technology/cersi-ai-shaping-safe-and-effective-ai-in-healthcare/ [Accessed 15 Nov. 2025]. Bell, S.K., Delbanco, T., Elmore, J.G., Fitzgerald, P.S., Fossa, A., Harcourt, K., Leveille, S.G., Payne, T.H., Stametz, R.A., Walker, J. and DesRoches, C.M. (2020). Frequency and Types of Patient-Reported Errors in Electronic Health Record Ambulatory Care Notes. JAMA Network Open , [online] 3(6), p.e205867. doi:https://doi.org/10.1001/jamanetworkopen.2020.5867. Bray, I. (2025). BMJ Group joins RADIANT to drive safe and effective digital health innovation - BMJ Group . [online] BMJ Group - Helping doctors make better decisions. Available at: https://bmjgroup.com/bmj-group-joins-radiant-to-drive-safe-and-effective-digital-health-innovation/ [Accessed 17 May 2025]. Brucato, C. (2023). Data Protection Policy . [online] Available at: https://www.gmc-uk.org/-/media/documents/data-protection-policy-gmc-v40.pdf [Accessed 16 May 2025]. Circlehealth.co. (2025). Circle Health - AI Agents for Recovery & Behavioral Programs . [online] Available at: https://circlehealth.co/ [Accessed 15 Nov. 2025]. CombineHealth. (2025). AI Revenue Cycle Management Software for Healthcare | CombineHealth . [online] Available at: https://www.combinehealth.ai/ [Accessed 15 Nov. 2025]. ContinuousCare (2020). All-in-One Virtual Practice with AI Scribe | ContinuousCare . [online] ContinuousCare. Available at: https://continuouscare.io/ [Accessed 15 Nov. 2025]. Corti.ai. (2016). Corti Assistant - AI Scribe for Clinical Documents . [online] Available at: https://assistant.corti.ai/ [Accessed 15 Nov. 2025]. DeepCura AI . (2025). DeepCura AI . [online] Available at: https://www.deepcura.com/ [Accessed 15 Nov. 2025]. Deepscribe.ai. (2025). DeepScribe AI Medical Scribe . [online] Available at: https://www.deepscribe.ai/ [Accessed 15 Nov. 2025]. Eleos Health. (2024). Automation and AI Solutions for Healthcare | Eleos Health . [online] Available at: https://eleos.health/ [Accessed 15 Nov. 2025]. Elion.health. (2025). Available at: https://elion.health/ [Accessed 15 Nov. 2025]. Empathia.ai. (2025). Empathia AI - AI Medical Scribe for Clinicians . [online] Available at: https://www.empathia.ai/ [Accessed 15 Nov. 2025]. England, N. (2023). NHS England» Digital clinical safety assurance . [online] England.nhs.uk. Available at: https://www.england.nhs.uk/long-read/digital-clinical-safety-assurance/#:~:text=What%20are%20DCB0129%20and%20DCB0160,of%20their%20health%20IT%20software. [Accessed 15 May 2025]. England, N. (2025). NHS England» Guidance on the use of AI-enabled ambient scribing products in health and care settings . [online] England.nhs.uk. Available at: https://www.england.nhs.uk/long-read/guidance-on-the-use-of-ai-enabled-ambient-scribing-products-in-health-and-care-settings/ [Accessed 16 May 2025]. Enzo.health. (2025). Home Health AI and Documentation Software | Enzo Health . [online] Available at: https://www.enzo.health/ [Accessed 15 Nov. 2025]. Epstein, J.A., Cofrancesco, J., Beach, M.C., Bertram, A., Hedian, H.F., Mixter, S., Yeh, H.-C. and Berkenblit, G. (2020). Effect of Outpatient Note Templates on Note Quality: NOTE (Notation Optimization through Template Engineering) Randomized Clinical Trial. Journal of General Internal Medicine , 36(3), pp.580–584. doi:https://doi.org/10.1007/s11606-020-06188-0. Evans, K., Papinniemi, A., Ploderer, B., Nicholson, V., Hindhaugh, T., Vuvan, V., Cowley, N., Tariq, A. and Thomson, H. (2025). Impact of using an AI scribe on clinical documentation and clinician-patient interactions in allied health private practice: perspectives of clinicians and patients. Musculoskeletal Science and Practice , [online] 78, p.103333. doi:https://doi.org/10.1016/j.msksp.2025.103333. expert.ai. (2025). Expert.ai – AI to Understand Data and Decide Better . [online] Available at: https://www.expert.ai/ [Accessed 15 Nov. 2025]. Fazakarley, C.A., Breen, M., Leeson, P., Thompson, B. and Williamson, V. (2023). Experiences of using artificial intelligence in healthcare: a qualitative study of UK clinician and key stakeholder perspectives. BMJ Open , [online] 13(12), p.e076950. doi:https://doi.org/10.1136/bmjopen-2023-076950. Fireflies.ai. (2025). Fireflies.ai | AI Teammate to Transcribe, Summarize, Analyze Meetings, Real Time AI Note Taker . [online] Available at: https://fireflies.ai/?utm_source=google&utm_medium=cpc&utm_campaign=brand_generic&utm_content=t1_eng&utm_term=fireflies%20ai&gad_source=1&gad_campaignid=23121070448&gbraid=0AAAAA-Os34pEfRBUfmRquvtPPQTlqlBT1&gclid=CjwKCAiAlMHIBhAcEiwAZhZBUoylRH1K2zGIn7l6gyRE7aG0Tq9QqKaIzEhdPnBRhBKps_KwbcorYRoC8wQQAvD_BwE [Accessed 15 Nov. 2025]. Getnudgeai.com. (2025). The Best AI Scribe and Coding Platform - Nudge AI . [online] Available at: https://getnudgeai.com/ [Accessed 15 Nov. 2025]. GOV.UK. (2014). Medical devices: software applications . [online] Available at: https://www.gov.uk/government/publications/medical-devices-software-applications-apps [Accessed 17 May 2025]. Guidance: Medical device stand-alone software including apps (including IVDMDs) v1.10f. (n.d.). Available at: https://assets.publishing.service.gov.uk/media/64a7d22d7a4c230013bba33c/Medical_device_stand-alone_software_including_apps__including_IVDMDs_.pdf. Hardianhealth.com. (2025). HaRi . [online] Available at: https://hari.hardianhealth.com/ [Accessed 15 Nov. 2025]. Health (2025). Fit for the future: 10 Year Health Plan for England - executive summary (accessible version) . [online] GOV.UK. Available at: https://www.gov.uk/government/publications/10-year-health-plan-for-england-fit-for-the-future/fit-for-the-future-10-year-health-plan-for-england-executive-summary [Accessed 15 Nov. 2025]. healthorbit.ai. (2025). Free AI Medical Scribe for Healthcare | HealthOrbit AI . [online] Available at: https://healthorbit.ai/ [Accessed 15 Nov. 2025]. Heidihealth.com. (2022). Heidi AI . [online] Available at: https://www.heidihealth.com/en-gb [Accessed 15 Nov. 2025]. Insighthealth.ai. (2025). Insight Health | AI Clinician Agents . [online] Available at: https://www.insighthealth.ai/ [Accessed 15 Nov. 2025]. JotPsych. (2024). JotPsych . [online] Available at: https://www.jotpsych.com/ [Accessed 15 Nov. 2025]. Kernberg, A., Gold, J.A. and Mohan, V. (2024). Using ChatGPT-4 to Create Structured Medical Notes From Audio Recordings of Physician-Patient Encounters: Comparative Study. Journal of Medical Internet Research , [online] 26, p.e54419. doi:https://doi.org/10.2196/54419. Kiwipen.com. (2025). Compliance | Kiwipen . [online] Available at: https://www.kiwipen.com/legal/compliance [Accessed 15 Nov. 2025]. Knowtex.ai. (2025). Knowtex - Voice AI Automated Clinical Workflows . [online] Available at: https://www.knowtex.ai/ [Accessed 15 Nov. 2025]. Kumpunen, S., Lobont, C., Garrard, L., Fisher, D., Lau, R., Fallica, G. and Fisher, R. (2025). How are GPs using AI? Insights from the front line . [online] Available at: https://www.nuffieldtrust.org.uk/sites/default/files/2025-12/How%20are%20GPs%20using%20AI_%20final.pdf [Accessed Dec. 2025]. Kwong, M.T., Stell, D. and Akinluyi, E. (2021). Medical Device Regulation from a Health Service Provider’s Perspective. Prosthesis , [online] 3(3), pp.261–266. doi:https://doi.org/10.3390/prosthesis3030025. Lanier, C., Dominicé Dao, M., Baer, D., Haller, D.M., Sommer, J. and Junod Perron, N. (2021). How Do Patients Want Us to Use the Computer During Medical Encounters?—A Discrete Choice Experiment Study. Journal of General Internal Medicine , 36(7), pp.1875–1882. doi:https://doi.org/10.1007/s11606-021-06753-1. Li, B., Crampton, N., Yeates, T., Lei, Y., Tian, X. and Truong, K.N. (2021). Automating Clinical Documentation with Digital Scribes: Understanding the Impact on Physicians. CHI . doi:https://doi.org/10.1145/3411764.3445172. Lyons, P.G., Rojas, J.C., Bewley, A.F., Malone, S.M. and Lekshmi Santhosh (2024). Validating the Physician Documentation Quality Instrument for Intensive Care Unit–Ward Transfer Notes. ATS Scholar , 5(2), pp.274–285. doi:https://doi.org/10.34197/ats-scholar.2023-0094oc. Lyrebirdhealth.com. (2025). Privacy & Security - Lyrebird Health . [online] Available at: https://www.lyrebirdhealth.com/uk/uk-compliance [Accessed 15 Nov. 2025]. Marianaai.com. (2024). MarianaAI - Your Supercharged AI Medical Scribe . [online] Available at: https://marianaai.com/ [Accessed 15 Nov. 2025]. Marino, F., Alby, F., Zucchermaglio, C. and Marilena Fatigante (2023). Digital technology in medical visits: a critical review of its impact on doctor-patient communication. Frontiers in Psychiatry , [online] 14. doi:https://doi.org/10.3389/fpsyt.2023.1226225. Maticinside.ai. (2022). Home . [online] Available at: https://maticinside.ai/ [Accessed 15 Nov. 2025]. Mdhub.ai. (2025). The Clinical AI Assistant for Behavioral Health Clinicians | mdhub . [online] Available at: https://www.mdhub.ai/ [Accessed 15 Nov. 2025]. MEDILIT - AI Medical Scribe (The most accurate). (2023). MEDILIT . [online] Available at: https://medilit.com/ [Accessed 15 Nov. 2025]. MedTech Europe. . IVDR & MDR Survey Results 2024. medtecheurope.org. (2024). Available at: https://www.gov.uk/government/publications/medical-devices-software-applications-apps [Accessed 1 Nov. 2025]. Mennella, C., Maniscalco, U., Pietro, G.D. and Esposito, M. (2024). Ethical and regulatory challenges of AI technologies in healthcare: A narrative review. Heliyon , [online] 10(4), pp.e26297–e26297. doi:https://doi.org/10.1016/j.heliyon.2024.e26297. Mentalyc.com. (2021). Mentalyc - AI Powered Notes & Insights for Deeper Therapy . [online] Available at: https://www.mentalyc.com/ [Accessed 15 Nov. 2025]. Meskó, B. and Topol, E.J. (2023). The imperative for regulatory oversight of large language models (or generative AI) in healthcare. npj Digital Medicine , [online] 6(1). doi:https://doi.org/10.1038/s41746-023-00873-0. Mhra.gov.uk. (2025). PARD . [online] Available at: https://pard.mhra.gov.uk/manufacturer-search/Microsoft%20DAX%20 [Accessed 15 Nov. 2025]. Moore, N. (2025). Compliance with NHS AI guidance and assurance standards - Lexacom . [online] Lexacom. Available at: https://www.lexacom.co.uk/compliance-with-nhs-ai-guidance-and-assurance-standards/ [Accessed 15 Nov. 2025]. Mutuohealth.com. (2025). AutoScribe | AI-Powered Medical Transcription and Note Generation . [online] Available at: https://www.mutuohealth.com/ [Accessed 15 Nov. 2025]. Nabla.com. (2025). Nabla · Enjoy care again . [online] Available at: https://www.nabla.com/ [Accessed 15 Nov. 2025]. National Data Guardian (2020). The Eight Caldicott Principles . [online] National Data Guardian, pp.1–2. Available at: https://assets.publishing.service.gov.uk/media/5fcf9b92d3bf7f5d0bb8bb13/Eight_Caldicott_Principles_08.12.20.pdf. NextGen Healthcare (2025). NextGen Healthcare . [online] NextGen. Available at: https://www.nextgen.com/ [Accessed 15 Nov. 2025]. Nguyen, A.D., White, S.J., Tse, T., Cartmill, J.A., Roger, P., Hatem, S. and Willcock, S.M. (2024). Communication during telemedicine consultations in general practice: perspectives from general practitioners and their patients. BMC Primary Care , 25(1). doi:https://doi.org/10.1186/s12875-024-02576-1. Nicholson, W., Gerke, S. and Cohen, I.G. (2024). Liability for use of artificial intelligence in medicine1. Edward Elgar Publishing eBooks , [online] pp.150–166. doi:https://doi.org/10.4337/9781802205657.ch09. Nora AI. (2025). Nora AI . [online] Available at: https://www.noramed.ai/ [Accessed 16 Nov. 2025]. Nuance Communications. (2025). Nuance Communications . [online] Available at: https://www.nuance.com/index.html [Accessed 15 Nov. 2025]. OmniMD. (2025). EHR, Billing & AI Solutions to Grow Your Practice - OmniMD . [online] Available at: https://omnimd.com/ [Accessed 15 Nov. 2025]. Orbdoc.com. (2025). OrbDoc - AI Medical Scribe for Physicians . [online] Available at: https://orbdoc.com/index.html [Accessed 15 Nov. 2025]. Oskrochi, Y., Roy-Highley, E., Grimes, K. and Shah, S. (2025). Digitally Unsafe? Digital health technology compliance with clinical safety standards within the NHS in England: a national cross-sectional study. (Preprint). Journal of Medical Internet Research . doi:https://doi.org/10.2196/80076. Pabau. (2025). The all-in-one Practice Management Software you need to run your Business . [online] Available at: https://pabau.com/ [Accessed 15 Nov. 2025]. Patientnotes.ai. (2025). Patient Notes . [online] Available at: https://patientnotes.ai/us [Accessed 15 Nov. 2025]. Patientnotes.app. (2025). PatientNotes - Let AI write your clinical notes for you. [online] Available at: https://www.patientnotes.app/auth/signup [Accessed 28 May 2025]. Playbackhealth.com. (2025). Ai Ambient Listening Scribe for Healthcare | Playback Health . [online] Available at: https://www.playbackhealth.com/ [Accessed 15 Nov. 2025]. Racgp.org.au. (2018). RACGP - Artificial intelligence (AI) scribes . [online] Available at: https://www.racgp.org.au/running-a-practice/technology/business-technology/artificial-intelligence-ai-scribes [Accessed 17 May 2025]. Reddy, S. (2024). Generative AI in healthcare: an implementation science informed translational path on application, integration and governance. Implementation Science , [online] 19(1). doi:https://doi.org/10.1186/s13012-024-01357-9. S10.ai. (2025). Best Ambient AI Medical Scribe for any EHR | S10.AI . [online] Available at: https://s10.ai/ [Accessed 15 Nov. 2025]. Scribe, A. (2025). Chartnote | AI Scribe, Medical Dictation, Snippets . [online] Chartnote.com. Available at: https://chartnote.com/ [Accessed 15 Nov. 2025]. ScribeAmerica. (2023). ScribeAmerica Commercial - ScribeAmerica . [online] Available at: https://www.scribeamerica.com/video/scribeamerica-commercial/ [Accessed 15 Nov. 2025]. Scribeberry (2025). Scribeberry | Never Chart Again . [online] Scribeberry. Available at: https://scribeberry.com/ [Accessed 15 Nov. 2025]. ScribeEMR . (2025). Best Virtual Medical Scribing Services | Remote Scribe Company . [online] Available at: https://www.scribeemr.com/ [Accessed 15 Nov. 2025]. Scribemd.ai. (2025). ScribeMD.ai - Your AI medical assistant . [online] Available at: https://www.scribemd.ai/ [Accessed 15 Nov. 2025]. ScribePT. (2025). The Best AI Scribe for Rehab Therapists l HIPAA Compliant . [online] Available at: https://www.scribept.com/ [Accessed 15 Nov. 2025]. Secinaro, S., Calandra, D., Secinaro, A., Muthurangu, V. and Biancone, P. (2021). The role of artificial intelligence in healthcare: a structured literature review. BMC Medical Informatics and Decision Making , [online] 21(1). Available at: https://link.springer.com/article/10.1186/s12911-021-01488-9. Service.gov.uk. (2023). Supplier Ambient Voice Technology (AVT) Self Certified Registry for the use of AI-Enabled Ambient Scribing Products in Health and Care Settings - Find a Tender . [online] Available at: https://www.find-tender.service.gov.uk/Notice/069369-2025?origin=SearchResults&p=138 [Accessed 15 Nov. 2025]. Seuren, L.M., Sakari Ilomäki, Evi Dalmaijer, Shaw, S.E. and Wyke (2024). Communication in Telehealth: A State-of-the-Art Literature Review of Conversation-Analytic Research. Research on language and social interaction , 57(1), pp.73–90. doi:https://doi.org/10.1080/08351813.2024.2305045. Simboalphus.com. (2025). SimboAlpus - Voice AI Assistant for Doctors for EMR Dictation . [online] Available at: https://secure.simboalphus.com/ [Accessed 15 Nov. 2025]. Sinnott, C., Moxey, J.M., Marjanovic, S., Leach, B., Hocking, L., Ball, S., Georgiadis, A., Lamé, G., Willars, J. and Dixon-Woods, M. (2021). Identifying how GPs spend their time and the obstacles they face: a mixed-methods study. British Journal of General Practice , p.BJGP.2021.0357. doi:https://doi.org/10.3399/bjgp.2021.0357. SOAPNoteAI (2025). Best AI SOAP Note Generator - Free Trial | SOAPNoteAI . [online] SOAPNoteAI.com. Available at: https://soapnoteai.com/ [Accessed 15 Nov. 2025]. Sollof, J. (2025). Accurx and Tandem Health roll out AI scribing tool across the NHS . [online] Digital Health. Available at: https://www.digitalhealth.net/2025/04/accurx-and-tandem-health-partner-to-roll-out-ai-scribing-in-the-nhs/ [Accessed 16 May 2025]. Staff Writers (2025). NHS drops ‘medical device’ bomb on AI scribes in UK - Medical Republic . [online] Medical Republic. Available at: https://www.medicalrepublic.com.au/nhs-drops-medical-device-bomb-on-ai-scribes-in-uk/117826 [Accessed 15 Nov. 2025]. Stokel-Walker, C. (2025). The ‘ambient scribe’ tools listening to and summarising your doctor-patient consultations. BMJ , [online] 389, p.r663. doi:https://doi.org/10.1136/bmj.r663. Sully.ai. (2025). AI For Doctors | Terms of Service - Sully . [online] Available at: https://www.sully.ai/data-processing-agreement [Accessed 15 Nov. 2025]. Swiftynote.com. (2025). SwiftyNote . [online] Available at: https://www.swiftynote.com/ [Accessed 15 Nov. 2025]. Tierney, A.A., Gayre, G., Hoberman, B., Mattern, B., Ballesca, M., Kipnis, P., Liu, V. and Lee, K. (2024). Ambient Artificial Intelligence Scribes to Alleviate the Burden of Clinical Documentation. NEJM Catalyst , [online] 5(3). doi:https://doi.org/10.1056/cat.23.0404. Tierney, A.A., Gayre, G., Hoberman, B., Mattern, B., Ballesca, M., Wilson, S.B., Castilla, K., Lau, C.S., Kipnis, P., Liu, V. and Lee, K. (2025). Ambient Artificial Intelligence Scribes: Learnings after 1 Year and over 2.5 Million Uses. NEJM Catalyst , [online] 6(5). doi:https://doi.org/10.1056/cat.25.0040. Tortus. (2025). TORTUS | Medical AI Healthcare Assistant | Doctor Approved . [online] Available at: https://tortus.ai/ [Accessed 15 Nov. 2025]. Trytwofold.com. (2025). Best AI Medical Scribe for Clinicians | Twofold Health . [online] Available at: https://www.trytwofold.com/ [Accessed 15 Nov. 2025]. Tzanou, M. (2020). Health Data Privacy under the GDPR . Routledge. Upheal.io. (2025). Upheal | AI progress notes . [online] Available at: https://www.upheal.io/ [Accessed 15 Nov. 2025]. Vero Scribe (2025). Vero Scribe . [online] Vero Scribe. Available at: https://secure.veroscribe.com/consultation [Accessed 15 Nov. 2025]. Additional Declarations The authors declare no competing interests. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8994700","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":598504875,"identity":"80c68c2b-7637-4216-ae6a-2ce277df06d3","order_by":0,"name":"Claire Dady","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABEklEQVRIiWNgGAWjYDACdgaGA0CKsYEhgYHhA48NWAQEDHBqYUbSwjhDJg0sQlALA0wLM4/NYcJa+JuZHx6uqGCQ7W/PMWDgyTmf2N/MwPjhB8NhY1xaJA6zGRw8c4bBeMaZNwYMEmduJ844zMAs2cNw2AyXFgNmBoODjW0MiQ03gLYY9txO3AB0mDQDw2Eb3FrYPxxs/MeQOB+kJfHfOZAW5t/4tfAAbWlgSNwA0nKA5wBICxvIFpwOkzjMU3Cw4ZiE8cYzz4AMnmTjGYcZ2yx7DNJxep+/vX3zx4YaG9l5x5M3Pv7DYwcMuubDN35UWBs24NIDtQyIOQwOQDjAOMITkciA/QFRykbBKBgFo2DkAQDO1VaozgNdIQAAAABJRU5ErkJggg==","orcid":"","institution":"University College London","correspondingAuthor":true,"prefix":"","firstName":"Claire","middleName":"","lastName":"Dady","suffix":""},{"id":598504876,"identity":"73ebdd98-e730-4a5d-884b-d1934a767604","order_by":1,"name":"Henry Potts","email":"","orcid":"","institution":"University College London","correspondingAuthor":false,"prefix":"","firstName":"Henry","middleName":"","lastName":"Potts","suffix":""},{"id":598504877,"identity":"2aa94f49-0016-4c7b-820a-5740a1bf69bb","order_by":2,"name":"Fiona Stevenson","email":"","orcid":"","institution":"University College London","correspondingAuthor":false,"prefix":"","firstName":"Fiona","middleName":"","lastName":"Stevenson","suffix":""},{"id":598504878,"identity":"1286f06c-5f25-4dc5-91f8-099f0ca8b28d","order_by":3,"name":"Patricia Schartau","email":"","orcid":"","institution":"University College London","correspondingAuthor":false,"prefix":"","firstName":"Patricia","middleName":"","lastName":"Schartau","suffix":""}],"badges":[],"createdAt":"2026-02-28 11:05:51","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-8994700/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8994700/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":104400552,"identity":"367f82f2-3e7d-4e67-811e-b4c03b46a45e","added_by":"auto","created_at":"2026-03-11 12:10:20","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1004182,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8994700/v1/495a6c8c-920f-4618-9f1b-86da7393dc25.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eTrust Me, I Might be a Medical Device: The Problem with AI Scribes\u003c/strong\u003e\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eRecent advances in large language models (LLMs) have driven the development of AI-based digital medical scribes, systems designed to capture clinician, patient interactions and generate structured documentation for inclusion in electronic health records. Depending on the vendor, these tools provide real-time transcription, automated summarisation, SNOMED coding, and the generation of formatted consultation notes or referral letters.\u003c/p\u003e \u003cp\u003eThe NHS Ten-Year Plan (2025) identifies artificial intelligence as a central enabler of digital transformation, explicitly referencing AI-enabled documentation assistants as a route toward a more data-driven and efficient healthcare system.\u003c/p\u003e \u003cp\u003eIn parallel, NHS England issued guidance in April 2025 confirming that AI scribes with specific functionalities may fall within the scope of the UK Medical Devices Regulations 2002, requiring registration with the Medicines and Healthcare products Regulatory Agency (MHRA) as software medical devices. In January 2026, NHS England introduced an AI-Enabled Ambient Voice Technology (AVT) Registry hoping to improve transparency and enable suppliers to self-certify compliance with NHS standards, including DTAC and MHRA requirements.\u003c/p\u003e \u003cp\u003ePersistent weaknesses in oversight across digital health technologies reinforce the need for such mechanisms. Many tools currently deployed within the NHS lack robust evidence of safety, benefit, or regulatory compliance, reflecting fragmented accountability and insufficient post-market evaluation (Oskrochi et al., 2025).\u003c/p\u003e \u003cp\u003eDespite these policy advances, the UK market for AI scribes remains uncoordinated. Many products are marketed directly to clinicians or small practices without demonstrable alignment to UK regulatory or clinical safety standards. This study presents a structured review of AI scribe products currently available to UK clinicians. It analyses publicly available vendor information and MHRA registration data to assess the transparency of regulatory claims and the challenges clinicians face in verifying compliance.\u003c/p\u003e \u003cp\u003eSpeech-to-text transcription tools have long been used in healthcare, but the latest generation of AI scribes marks a significant shift in capability. According to NHS England\u0026rsquo;s April 2025 guidance, newer tools leveraging LLMs to summarise consultations are no longer considered passive documentation aids. Depending on their functionality, they may fall within the scope of the UK Medical Devices Regulations 2002 (NHS England, 2025b). The Medicines and Healthcare products Regulatory Agency had previously, on 3 February 2025, concluded that scribes in mental health likely fall within this remit.\u003c/p\u003e \u003cp\u003eThe NHS 10-Year Health Plan reinforces this evolution, positioning ambient AI (AVT) and digital assistants as central to the NHS\u0026rsquo;s digital-transformation strategy. Pilot evaluations cited in the plan report that AI documentation tools have reduced administrative workload by over 50%, enabling staff to focus more on patient care and improving service throughput (NHS England, 2025a).\u003c/p\u003e \u003cp\u003eAI tools are increasingly present in UK primary care, with around one-third of GPs already using them, yet meaningful impact is not assured. Adoption remains uneven, shaped by differing local policies in which some ICBs encourage cautious piloting while others promote more proactive use of approved systems (Kumpunen et al., 2025).\u003c/p\u003e \u003cp\u003eHowever, these innovations raise complex regulatory and ethical challenges. The MHRA defines software as a medical device as any software intended for a medical purpose, including diagnosis, treatment, or prognosis. Even if a developer does not explicitly claim medical use, a tool\u0026rsquo;s interface, marketing, or deployment context can bring it within regulatory scope. Features such as automated summarisation, decision support, or clinical coding particularly when not easily verifiable by clinicians, increase the likelihood of classification as a medical device (MHRA, v1.10f).\u003c/p\u003e \u003cp\u003eOskrochi et al. (2025) identified persistent deficiencies in compliance, governance, and post-market monitoring across a wide range of digital health tools currently in use. Their analysis demonstrated that many technologies are implemented without clear evidence of safety, benefit, or regulatory alignment, and that responsibility for compliance remains fragmented across multiple agencies. The authors concluded that NHS oversight mechanisms have not kept pace with the rapid evolution of digital innovation, resulting in gaps in accountability and transparency. In response to these challenges, NHS England has now introduced a self-certified AVT supplier registry, intended to act as a national reference point for organisations and clinicians seeking assurance on regulatory and governance alignment \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://transform.england.nhs.uk/digitise-connect-transform/digitising-the-frontline/ambient-voice-technology-self-certified-supplier-registry/\u003c/span\u003e\u003cspan address=\"https://transform.england.nhs.uk/digitise-connect-transform/digitising-the-frontline/ambient-voice-technology-self-certified-supplier-registry/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. The registry enables suppliers to declare compliance with NHS requirements, including DTAC, clinical safety standards, and (where applicable) MHRA medical device registration. While this represents an important structural intervention, the registry relies on self-certification rather than independent verification, and its effectiveness will depend on supplier participation, accuracy of declarations, and visibility to end-users. At present, there is limited evidence regarding clinician awareness of the registry or its integration into routine procurement, governance, or audit processes.\u003c/p\u003e \u003cp\u003eRegulatory classification carries significant implications. Products meeting the definition of a medical device must be registered with the MHRA, comply with UK Conformity Assessed (UKCA) or CE-marking requirements, and adhere to NHS clinical-safety standards (DCB0129 and DCB0160). These obligations may present barriers for smaller developers or overseas vendors unfamiliar with UK-specific regulation. The 10-Year Plan acknowledges these challenges, committing to a \u0026ldquo;faster, predictable, and risk-proportionate\u0026rdquo; regulatory framework for AI and software-as-a-medical-device, developed jointly by the MHRA and the Department for Science, Innovation and Technology (NHS England, 2025a).\u003c/p\u003e \u003cp\u003eAlthough the statutory MHRA registration fee for software devices is modest, currently \u0026pound;261 (rising to \u0026pound;300 annually in 2026), the true cost of achieving full regulatory compliance is substantial. When accounting for engineering design, risk management, regulatory documentation, clinical validation, and conformity assessment, the average development cost for a Class II medical device (including digital and software-only systems) ranges between \u0026pound;1.6\u0026nbsp;million and \u0026pound;4\u0026nbsp;million (MedTech Europe, 2024), albeit software-only devices typically occupy the lower end of this range.\u003c/p\u003e \u003cp\u003eThere is a high turnover of small providers in the emerging digital-scribe (AVT) market. The advent of accessible LLM frameworks and open developer APIs has lowered the barrier to entry, enabling start-ups and independent developers to produce AI documentation tools with limited resources. Many of these vendors market directly to clinicians and small practices via web platforms, promoting their tools as workflow aids rather than regulated medical devices. However, the cost and complexity of full Class II certification, including quality-management systems and clinical-performance evidence, may be prohibitive for all but the largest or most established suppliers.\u003c/p\u003e \u003cp\u003eProfessional responsibility nevertheless remains with the clinician, who must review and verify AI-generated outputs before they are entered into the patient record. Failure to identify inaccuracies could expose clinicians to liability if harm occurs, even when errors originate from the AI system (Nicholson et al., 2024). Automation bias where clinicians over-rely on plausible AI outputs has been recognised as an emerging safety concern (Abdelwanis et al., 2024). Further governance questions surround patient consent and data transparency. While some providers verbally inform patients when AI scribes are used, consent models vary. Under UK GDPR and the Data Protection Act 2018, patients have the right to understand how their data are processed, and professional guidance from the GMC and National Data Guardian emphasises the need for transparency. Although the NHS envisions a unified, digital-first health record, it is not yet standard practice to mark AI-generated documentation within patient records. This omission has implications for accountability, auditability, and trust (Mennella et al., 2024).\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003eThis study aimed to identify AI-powered digital medical scribes available to clinicians in the United Kingdom and assess the transparency of their regulatory status. The analysis focused on tools that are marketed as using AI to transcribe or summarise patient consultations and are promoted for use in healthcare settings.\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eEligibility Criteria\u003c/h2\u003e \u003cp\u003eTo be included in the review, a product had to:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eBe marketed as an AI-powered digital scribe or consultation assistant in a health care setting including mental health/therapy;\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eOffer transcription, summarisation, or clinical documentation support based on patient\u0026ndash;clinician interaction;\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eTools offering only basic dictation or voice recognition (e.g. standalone speech-to-text apps) were excluded unless they also included generative summarisation or structured documentation functionality.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eSearch Strategy\u003c/h3\u003e\n\u003cp\u003eThe search was conducted between 1\u0026ndash;1\u003cem\u003e5\u003c/em\u003e Nov 2025 using a combination of formal and informal sources to identify products available to UK clinicians. This included:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eStructured web searches (Google) using terms such as \u0026ldquo;AI scribe UK,\u0026rdquo; \u0026ldquo;digital medical scribe NHS,\u0026rdquo; \u0026ldquo;ambient documentation tool,\u0026rdquo; and \u0026ldquo;LLM scribe healthcare\u0026rdquo;;\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eNHS procurement websites, innovation portals, and accelerator programme listings;\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003ePublicly accessible regulatory databases including the MHRA Public Access Medical Device Registration System;\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eU.S healthcare technology data base - Elion;\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eCompany websites, marketing materials, press releases, and white papers;\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eProfessional discussion forums, including Reddit threads and X posts;\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eLinkedIn company pages, employee posts, and industry analyst commentary.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eThis approach generated a list of 121 AI scribe tools potentially available to UK clinicians. These included both UK-developed and international tools, some of which are in active NHS deployment or pilot phases.\u003c/p\u003e\n\u003ch3\u003eData Extraction and Review\u003c/h3\u003e\n\u003cp\u003eFor each tool, data was manually extracted from public sources, including:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eProduct features (e.g. transcription, summarisation, coding);\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eRegulatory status (e.g. MHRA registration);\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eIntegration claims with UK EHR systems and NHS compliance (e.g. EMIS, SystmOne);\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eData handling practices (e.g. patient data storage, audio retention, anonymisation);\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003ePatient consent processes and governance indicators, if any (e.g. documentation of AI use, consent procedures, audit trails).\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eWhere regulatory status was unclear or unverifiable via official databases, this was recorded as a key finding.\u003c/p\u003e\n\u003ch3\u003eLimitations\u003c/h3\u003e\n\u003cp\u003eThis review has several important limitations. First, data collection was completed prior to the publication of the NHS England Ambient Voice Technology (AVT) self-certified supplier registry in January 2026. As a result, the analysis does not account for supplier participation in, or compliance declarations made through, this registry. The findings therefore represent a snapshot of the market immediately before the introduction of a national assurance mechanism and may over-estimate the current degree of opacity in regulatory signalling.\u003c/p\u003e \u003cp\u003eThe MHRA registration database is not searchable by product function, AI capability, or clinical application, which makes it difficult to confirm whether a product is in scope or appropriately registered.\u003c/p\u003e \u003cp\u003eSome tools may be undergoing registration, or may self-certify as Class I devices under UK Medical Devices Regulations 2002, which does not require independent third-party assessment. As stated in MHRA guidance:\u003c/p\u003e \u003cp\u003e \u003cem\u003e\u0026ldquo;Registration of medical devices with the MHRA... does not represent any form of accreditation, certification, approval or endorsement by the MHRA. Therefore, it is not permitted to make any claims to this effect\u0026hellip;\u0026rdquo;\u003c/em\u003e (MHRA, 2023).\u003c/p\u003e \u003cp\u003e As such, the presence of a tool on the register should not be interpreted as confirmation of compliance with NHS safety, performance, or clinical governance expectations. Additionally, several vendors do not publish regulatory documentation or safety information on their websites, further limiting the ability to verify claims.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eOverview\u003c/h2\u003e \u003cp\u003eA total of 118 AI-enabled ambient scribing (AVT) products were identified as being marketed to, or accessible by, UK clinicians. These include UK-based and international suppliers offering capabilities such as ambient voice capture, real-time transcription, generative summarisation, structured coding, and automated referral generation. Claims of regulatory compliance, NHS readiness, and data-handling transparency varied widely.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"No\" id=\"Taba\" border=\"1\"\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003e Scribe Name\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMHRA Device Registration\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eUK Access Availability\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eCompliance Notes\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAbridge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePrivacy policy legal, no direct GDPR citation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAccel eq\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSome Accel privacy docs claim GDPR, check clinical context\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAesthetics 360\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUK sector guidance, no direct product claim\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eaide health\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes (Class 1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNHS DCB0129/0160 clinical safety standards, GDPR data protection.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAiva\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, privacy statement, EU/UK data rights\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAmbience Scribe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNHS-required GDPR, needs direct confirmation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAndy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, security audits\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAnima/ Annie (UK)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes (class 1 claimed), could not find on pard\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUK GDPR,Use in nhs reported, DPA, DTAC, DSPT, DCB0129\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAugmedix (USA)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHIPAA, de-identification, US focus\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAugnito (UK/IND)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes (class 1) (ScribeTech)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNHS dictation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUK GDPR, NHS, local hosting, security, DPA\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAura AI Scribe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, consent\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAutoMynd\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, DPO, consent, EU hosting\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAvo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, CCPA, anonymization, audit\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAvodahmed\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, patient consent, subject rights, industry audits\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBastionGPT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, UK/EU legal review\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBells ai\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, PIPEDA, UK data controls\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBlueprint\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eReferences OpenAI's privacy; not product-level GDPR\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCare cortex\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, ISO, DPO, subject rights\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCarePilot\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, UK DPA, Aus Privacy Act, DPO\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eChartnote\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, PIPEDA, SOC2, staff training, logging\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCircleHealth AI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHIPAA, Joint Commission, lacks GDPR\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eClinicminds\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, PIPEDA, privacy, consent management\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCommure Ambient AI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, end-to-end encryption\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eContinuous care\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, EU/UK hosting, right to be forgotten\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eConveyor AI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, DPA/EU/UK data addendum\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCorti\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, ISO27001, SOC2, medical device certs\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDeepCura AI (USA/CAN)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, DPO, ongoing audits\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDeepScribe (USA)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eUnclear\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHIPAA, SOC2 only\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDeliberate ai\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNo public policy, GDPR not found\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDiktamen\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eno\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eDTAC, DCB0129, DSPT, Cyber esentials, ISO 27001, ISO 9001, ISO14001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDragon Medical One (USA)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNHS use claimed\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAzure UK/EU hosting, ISO27001, DCB0129, NHS compliance\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEleos\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUK privacy, GDPR/UK GDPR, subject rights, DPO, audits\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEmpathia ai\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, ISO27001, privacy basis\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEnzohealth\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, ISO27001, SOC2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eExpert Ai\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR for Cogito, DPO support, EU bases only for that prod.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEzy scribe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNo public GDPR documentation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFireflies ai\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, SOC2, HIPAA, US hosting, enterprise EU option\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFreed AI (USA)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, HITECH, audits, US/EU privacy\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGlass health\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHIPAA, US-centric only\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHalo scribe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNo public GDPR evidence\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHealth orbit\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes, NHS use claimed\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, UK/EU data handling, privacy framework\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHeidi Health (AUS)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes (Class 1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes, NHS use claimed\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, NHS, ISO27001, UK hosting\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eInnovaccer InScribe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, Trust Center, EU/UK\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eInsight aura AI scribe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, ISO27001, UK Cyber Essentials\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIscribe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNo (not supported)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHIPAA-only focus, no GDPR evidence\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eJessica by Combine\u003c/p\u003e \u003cp\u003eHealth\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHIPAA, CMS, lacking explicit GDPR\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eJotpsych\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2 Type 2, audit, deletion\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKiwipen (UK)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, ICO, Cyber Essentials, DSPT, DCB0129, DTAC\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKnowtex\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHIPPA, SOC 2, GDPR, iso 27001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLexacom Scribe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes States class one on website unable to find on pard\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, ISO, UK DPA, DPO, EU/UK data controls\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLucus ai\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, UK data residency\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLyrebird Health (AUS)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes (Class 1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes, NHS use claimed\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, NHS DTAC, DCB0129, HIPAA, Australia Privacy Act\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMariana ai\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHIPAA only, not GDPR for healthcare\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMarvix\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, SOC2, EEA data, subject rights, breach notification\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMdhub\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, security audits\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMediary\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNo privacy or public GDPR evidence, not AVT scribe\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMEDILIT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, product claim\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMediSync\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, Texas regional, Elion Health confirmed\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMentalyc\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, PHIPA, PIPEDA, rights, audit\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMircosoft DAX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes website states class 1 could not find on pard\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, HITRUST, SOC2, SCCs, EU/UK hosting\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMutuo AutoScribe (CAN)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, PHIPA, PIPEDA, UK/EU support\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNabla Copilot (FRA/USA)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, ISO27001, EU/US hosting\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNextGen Ambient Assist\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, Elion Health review\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNextvisit AI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNot GDPR compliant; US only, not for UK/EU (privacy stmt)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNora note\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, privacy, consent, UK/EU legal\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNudge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, minimization, data rights, consent\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOmniMD AI Medical Scribe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, CCPA, SOC2, audits, EU/UK controls\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOneChart\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNo public GDPR evidence or policy\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOneChart (CAN)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNo public GDPR evidence or policy\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOneLine Health\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eno\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, PIPEDA, NHS, privacy/erasure\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOpus\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, product privacy solutions\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOracle Clinical Digital Asst.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, DPA, SCC, EU privacy safeguards\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOrbDoc\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, ISO27001, NHS DSPT, UK/EU\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePabau - Echo AI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, consent/audit features\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePatientNotes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, UK DPA, Aus Privacy Act, audit logs\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePlayback health\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, explicit EU controls\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePlena Health AI Scribe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, UK/EU privacy controls\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGiiq\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, staff training, privacy, subject rights\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRegard\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHIPAA, SOC2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eReteta\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNo public GDPR privacy policy available\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRevmax\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, legal processing EU/UK, DPA, data rights\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRiver Records\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, UK DPA, UK/EU data, subject rights\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eS10.AI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, ISO, PIPEDA, EU hosting, DPA available\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSayvant\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, audits, EU/UK subject rights\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eScribble\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eError 403 forbidden message\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePrivacy error, possibly inaccessible in UK\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eScribe america speke\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHIPAA, US-centric only\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eScribeberry\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, ISO, audits, Canadian compliance\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003escribematic\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, ISO, CCPA, NHS guides, UK/EU privacy\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eScribeMD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, EU/UK subject rights\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eScribePT (USA)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUS privacy only\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eScribeRyte\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, ISO27001, privacy controls\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eShine AI beam health\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, independent audits, privacy policies\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSimboAlphus\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSome Symbolab legal basis, not healthcare certain\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSKRIBER\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eData protection, legal compliance\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSoap suds\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNo GDPR for health context\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSOAPNoteAI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, Cyber Essentials, ISO, PIPEDA\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSolventum fluency direct\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eno\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, TX-RAMP, Cyber Essentials, ISO 27001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSpeak matic\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eUnclear\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNo explicit GDPR found\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSplose\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eno\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, UK GDPR, ISO 27001, Health data encryption, SOC2, SSO/2FA\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSporo AI Scribe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, UK Cyber Essentials, NHS DSPT, ISO\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSuki AI (USA)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, anonymization, security, user consent\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSully u.s\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eUnclear\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, ISO 27001, SOC2, EEA/UK compliance\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSupanote AI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, PHIPA, PIPEDA, consent, audits\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSwiftyNote\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eno\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eUnclear. Code never sent for log in after request\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, AWS, privacy/security\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT-Pro\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNHS DCB0129/0160, DTAC, DSPT\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTali AI (CAN)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHIPAA, SOC2, PHIPA/PIPA, lacks explicit GDPR\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTandem\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes (Class1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes NHS use\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUK GDPR, NHSE, DPA, NHS toolkit, DPIA, CyberEssentials\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTheraPulse\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, Elion Health confirmed\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTortus (UK)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes (Class 1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes NHS use\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, CyberEssentials Plus, DTAC, DSPT, DCB0129, NHS\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTriple time\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNo GDPR evidence found in public docs\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTwofold Health (USA)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHIPAA, no public GDPR evidence\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUpheal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, PHIPA, PIPEDA, UK/EU hosting\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVeradigm Ambient Scribe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, EU/UK compliance\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVero\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, ISO27001, SOC2 Type II, HIPAA, external audits\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVidera Health\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eContact required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, privacy policy EU/UK\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWavo Health (CAN)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, SOC2, no audio storage, UK hosting\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eZanda\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGDPR, HIPAA, ISO, UK DPA, DPO, EU/UK data controls\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003eRegulatory verification (MHRA)\u003c/h2\u003e \u003cp\u003eManual verification against the Medicines and Healthcare products Regulatory Agency (MHRA) Public Access Registration Database (PARD) \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://pard.mhra.gov.uk/\u003c/span\u003e\u003cspan address=\"https://pard.mhra.gov.uk/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e identified nine of 118 AI-enabled AVT products (8%) that claimed Class I medical device status. Of these nine products, six (5% of the identified AVT products) could be definitively located on PARD under the same or a clearly linked manufacturer or legal entity name at the time of review. For the remaining three products (3% of all AVT products), no corresponding PARD entry could be identified under the marketed product name. While these products may be registered under alternate legal entities or parent companies, this could not be confirmed using publicly available records.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePublicly stated and verifiable MHRA regulatory status of AI-enabled ambient voice technology products marketed to UK clinicians (N\u0026thinsp;=\u0026thinsp;118)\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRegulatory status\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCount\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e% of total\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eClaims MHRA Class I medical device status\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVerified on MHRA PARD under same or clearly linked legal entity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eClaimed Class I status but not verifiable on PARD under marketed name\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo claim of MHRA Class I status\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e109\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e92\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cem\u003eVerification was defined as a clear and defensible link between the marketed product and a registered legal entity on the MHRA Public Access Registration Database (PARD). Products not locatable under the marketed name may be registered under alternative legal or trading names.\u003c/em\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eData protection and GDPR compliance\u003c/h2\u003e \u003cp\u003eOf the 118 AVT products reviewed, 88 (75%) provided an explicit statement of compliance with the General Data Protection Regulation (GDPR) or UK data protection legislation on their websites or in associated documentation. Only a minority of these products linked supporting evidence, such as Data Protection Impact Assessments or references to specific UK GDPR provisions. The remaining products either provided general statements relating to privacy or security without explicit reference to GDPR or did not provide sufficient publicly available information to determine compliance.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePublicly stated GDPR and data protection compliance among AI-enabled ambient voice technology products (N\u0026thinsp;=\u0026thinsp;118)\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eData protection status\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCount\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e% of total\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eExplicit GDPR or UK data protection compliance statement\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e88\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e75\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGeneric privacy or security statements without explicit GDPR reference\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e19\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo clear data protection information identified\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cem\u003eExplicit compliance was defined as a direct reference to GDPR or UK data protection legislation in publicly available product documentation or websites.\u003c/em\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eUK access and deployment\u003c/h2\u003e \u003cp\u003eMost AVT products demonstrated some degree of potential accessibility to the UK market, although access pathways varied considerably. Forty-nine products (42%) required direct vendor contact or a product demonstration prior to access, indicating possible UK availability but limited public transparency. Forty-three products (36%) were clearly accessible to UK users or reported active or planned deployment within UK or NHS settings. Fifteen products (13%) explicitly stated that they were not available to UK users or were not accesessible from a UK IP address, while 11 products (9%) had unclear or indeterminate access status due to insufficient or inconclusive publicly available information.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eUK access and deployment characteristics of AI-enabled ambient voice technology products (N\u0026thinsp;=\u0026thinsp;118)\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAccess category\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCount\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e% of total\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eInterpretation\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eContact or demonstration required\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e49\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e42\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePotential UK availability but limited public transparency\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAccessible to UK users or NHS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e36\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eClear UK access or evidence of NHS use\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNot accessible to UK users\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eExplicit geographical or market restriction\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUnclear or indeterminate\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eInsufficient or inconclusive public information\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eA substantial proportion of AVT products referenced current or intended use within the NHS, commonly in the context of pilot projects, evaluations, or integrations with established clinical systems such as EMIS or SystmOne. Notably, all products that claimed MHRA Class I medical device status also reported current or planned use within NHS settings.\u003c/p\u003e \u003cp\u003eSeveral products without verifiable Class I registration displayed NHS logos or referenced NHS partnerships on their websites. While such references suggest market intent, they were not consistently supported by verifiable evidence of formal contracts or regulatory approval. In addition, a number of products appeared to be primarily aligned with international regulatory frameworks, particularly those applicable in the United States, rather than explicitly addressing UK-specific regulatory and governance requirements.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eObserved Trends and Common Gaps\u003c/h2\u003e \u003cp\u003eBrand\u0026ndash;manufacturer discrepancies in MHRA listings can limit traceability and slow due diligence checks. In addition, cross-border data processing arrangements are often not clearly specified on supplier websites, including whether appropriate UK safeguards are in place when data are stored or processed outside the UK. Marketing assurances such as claims of being \u0026ldquo;secure\u0026rdquo; or \u0026ldquo;HIPAA compliant\u0026rdquo; are frequently presented without sufficient detail to meet NHS assurance requirements. These results reinforce the case for a central NHS mechanism such as the AVT Registry to provide verifiable evidence, reduce duplication of effort across ICBs, and support consistent, safe procurement decisions.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003e This review highlights a significant gap between the regulatory expectations outlined in NHS and MHRA guidance and the transparency of AI scribe tools currently available to UK clinicians. Despite the 2025 NHS statement that high-functionality AI scribes may fall under the UK Medical Devices Regulations 2002, only a minority of tools reviewed had verifiable MHRA registration, UKCA/CE certification, or published clinical safety documentation.\u003c/p\u003e \u003cp\u003eFor frontline clinicians, this creates uncertainty and potential professional risk. AI scribes that generate summaries, add clinical codes, or format notes for upload into patient records are not merely passive transcription tools. If a tool meets the legal definition of a medical device, clinicians using unregistered versions may bear liability in the event of harm, particularly if due diligence around regulation and safety was not exercised. Moreover, automation bias may lead clinicians to over-rely on AI-generated notes, especially under time pressure or when the output appears plausible.\u003c/p\u003e \u003cp\u003eYet the appeal of AI scribes is undeniable. These tools offer benefits to clinicians and patients alike, particularly for healthcare professionals with high administrative burdens. In the context of increasingly time-constrained consultation models, tools that can reduce cognitive load and streamline note-taking hold real promise for improving care delivery (Tierney et al., 2025).\u003c/p\u003e \u003cp\u003eThe challenge lies in the pace of development. Rapid advances in generative AI, including the integration of large language models such as GPT-4, have outstripped the capacity of existing regulatory systems to adapt. Today\u0026rsquo;s AI scribes go far beyond speech-to-text, generating structured summaries, coded entries, and even referral letters within seconds (Sollof, 2025). For clinicians, this raises the question of how they can critically evaluate and safely deploy tools whose regulatory status, audit mechanisms, and error-handling processes remain unclear.\u003c/p\u003e \u003cp\u003eThis concern becomes especially pressing when AI scribes are embedded directly into electronic patient record (EPR) systems. Errors in transcription, summarisation, or clinical coding may be silently propagated across systems, leading to downstream risks. As more patients gain access to their health records, inaccuracies could generate confusion, complaints, or requests for correction. Misattributed diagnoses or incorrect codes could directly contribute to patient harm and may also create new burdens for data protection officers as well as compromising clinician\u0026ndash;patient trust.\u003c/p\u003e \u003cp\u003eThese findings also suggest that while the majority of AVT scribes have some pathway to the UK market, accessibility and verifiable compliance remain inconsistent. Nearly half of all identified tools require vendor engagement in the form of an online demonstration and discussion with a sales agent before access is granted. Fewer than ten could demonstrate confirmed medical device registration. This suggests a market still in an early stage of regulatory maturity, where international vendors are open to UK adoption but have not fully aligned with UK-specific safety, data governance, or registration standards this does not stop them trying to sell into the UK market.\u003c/p\u003e \u003cp\u003eVendor engagement may, however, serve as a modest safeguard. The need to arrange a demonstration or contact a sales team introduces a degree of friction that could prevent clinicians from adopting tools without adequate due diligence or confirmation of GDPR compliance. In that sense, the \u0026ldquo;demo/contact required\u0026rdquo; model may help mitigate inappropriate or premature use. Nevertheless, discussions with several US-based vendors indicated a strong interest in selling to the UK market despite limited understanding of UK regulatory expectations. This lack of awareness could present risks for clinicians who assume these products meet local compliance standards when, in fact, they may only adhere to US frameworks such as HIPAA (Health Insurance Portability and Accountability Act).\u003c/p\u003e \u003cp\u003eIn parallel, the NHS would benefit from a national audit framework for AI-generated documentation. Standardised approaches are needed to determine how frequently outputs should be reviewed, what constitutes an acceptable error threshold, and how the role of clinical safety officers should evolve. Without such benchmarks, quality assurance will remain fragmented and reactive.\u003c/p\u003e \u003cp\u003eThe introduction of the NHS England AVT self-certified registry represents a potentially important inflection point in this landscape. If widely adopted and appropriately maintained, the registry could reduce duplication of due diligence across integrated care boards, support more consistent procurement decisions, and provide clinicians with a clearer route to verifying baseline compliance. However, for this potential to be realised, the registry must be visible, trusted, and actively embedded within NHS digital governance workflows. Without proactive communication, integration into existing NHS systems, or linkage to professional guidance, there is a risk that the registry remains peripheral, particularly as commercial vendors continue to market directly to clinicians outside formal procurement pathways.\u003c/p\u003e \u003cp\u003eFinally, while initiatives such as RADIANT (Regulatory and Innovation Network for AI Technologies) and CERSI AI (Centre of Excellence for Regulatory Science \u0026amp; Innovation in AI) are emerging to address the governance gap (Bray, 2025), their impact will depend on strong inter-agency coordination, visible national leadership and funding. A transparent, central register of compliant tools, such as the planned NHS AVT Registry, would provide clinicians and organisations with a trusted reference point. In the absence of such infrastructure, responsibility continues to fall on individual users to assess regulatory risk, a task that remains difficult in the current environment.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eAI digital scribes offer considerable promise in reducing documentation burden, improving clinical workflows, and enhancing the quality of records. However, this review has shown that despite clear NHS guidance, the current regulatory landscape lacks transparency and consistency. Many tools marketed to UK clinicians do not clearly disclose their regulatory status.\u003c/p\u003e\n\u003cp\u003eThe recent introduction of a national NHS self-certified AVT supplier registry is a welcome and necessary development, signalling recognition of the governance gap identified in this study. However, its late arrival relative to market growth, reliance on self-certification, and limited visibility underscore the need for continued scrutiny. To support safe, accountable adoption of AI scribes, registry participation must be coupled with clearer guidance for clinicians, improved discoverability, and alignment with audit and clinical safety processes. Without these measures, responsibility for managing regulatory risk will continue to fall on individual users, undermining both patient safety and clinician confidence in emerging AI documentation technologies.\u003c/p\u003e\n\u003cp\u003eWithout such action, the NHS risks entering a period of unchecked adoption, where the benefits of AI scribes may be undermined by avoidable risks to patients, clinicians, and healthcare systems alike.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eConflicts of Interest/Funding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eHWWP is partly funded as part of the RADIANT CERSI by the Medical Research Council, Innovate UK, Medicines and Healthcare products Regulatory Agency, and Office for Life Sciences. Views presented are the authors and do not represent those of the funding organisations. HWWP provides or has provided consultancy on digital health evaluation for Flo Health Inc. and Thrive Therapeutic Software Ltd. He has PhD students in the field employed by or previously employed by, and with fees paid by or previously paid by AstraZeneca, Patients Know Best, and BetterPoints Ltd.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAbdelwanis, M., Alarafati, H.K., Tammam, M.M.S. and Simsekler, M.C.E. (2024). Exploring the risks of automation bias in healthcare artificial intelligence applications: A Bowtie analysis. \u003cem\u003eJournal of Safety Science and Resilience\u003c/em\u003e, [online] 5(4), pp.460\u0026ndash;469. doi:https://doi.org/10.1016/j.jnlssr.2024.06.001.\u003c/li\u003e\n\u003cli\u003eAbridge.com. (2025). \u003cem\u003eGenerative AI for Clinical Conversations | Abridge\u003c/em\u003e. [online] Available at: https://www.abridge.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eAccurx.com. (2022). \u003cem\u003eAccurx Scribe: Privacy and Security FAQs | Accurx Help Centre\u003c/em\u003e. [online] Available at: https://support.accurx.com/en/articles/11156712-accurx-scribe-privacy-and-security-faqs [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eAesthetics360.com. (2025). \u003cem\u003eaesthetics360\u003c/em\u003e. [online] Available at: https://aesthetics360.com/ [Accessed 16 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eAI Medical Scribe for Clinicians | Wavo Health. (2025). \u003cem\u003eAI Medical Scribe for Clinicians | Wavo Health\u003c/em\u003e. [online] Available at: https://wavo.health/en [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eAI Medical Scribe for Healthcare \u0026amp; Medicine | Skriber. (2025). \u003cem\u003eHome - AI Medical Scribe for Healthcare \u0026amp; Medicine | Skriber\u003c/em\u003e. [online] Available at: https://skriber.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eAide.health. (2018). \u003cem\u003eAide for Clinicians \u0026mdash; Rebuilding chronic disease care with AI\u003c/em\u003e. [online] Available at: https://aide.health/products/aide [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eAiscribee.com. (2025). \u003cem\u003eHome | AI Medical Scribe \u0026amp; EMR Assistant | ScriBee\u003c/em\u003e. [online] Available at: https://aiscribee.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eAli, M., Fantesini, A., Morcella, M.T., Ibba, S., Gennaro D\u0026apos;Anna, Fazzini, D. and Papa, S. (2023). Adoption of AI in Oncological Imaging: Ethical, Regulatory, and Medical-Legal Challenges. \u003cem\u003eCritical Reviews\u003csup\u003eTM\u003c/sup\u003e in Oncogenesis\u003c/em\u003e, [online] 29(2), pp.29\u0026ndash;35. doi:https://doi.org/10.1615/critrevoncog.2023050584.\u003c/li\u003e\n\u003cli\u003eAmbiencehealthcare.com. (2025). \u003cem\u003eAmbience Healthcare\u003c/em\u003e. [online] Available at: https://www.ambiencehealthcare.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eAnaiva.io. (2025). \u003cem\u003eAI Medical Scribe Software | Automated Clinical Documentation | Aiva\u003c/em\u003e. [online] Available at: https://www.anaiva.io/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eAnimahealth.com. (2025). \u003cem\u003eAnima | Anima - Full Compliance with NHS England AVT Requirements\u003c/em\u003e. [online] Available at: https://www.animahealth.com/blog/anima-full-compliance-with-nhs-england-avt-requirements [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eAppliedAI - Opus. (2025). \u003cem\u003eOpus - AI Workflow and Business Automation Platform | Opus\u003c/em\u003e. [online] Available at: https://www.opus.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eAugmedix.com. (2025). \u003cem\u003eTerms of Agreement | Augmedix\u003c/em\u003e. [online] Available at: https://www.augmedix.com/terms-of-agreement/current [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eAugnito.ai. (2025). \u003cem\u003eAugnito | Enhance Clinical Documentation with Voice AI\u003c/em\u003e. [online] Available at: https://augnito.ai/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eAutomynd.com. (2024). \u003cem\u003eOrchestrating AI for Home-Based Care Operations\u003c/em\u003e. [online] Available at: https://www.automynd.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eAvaamo. (2021). \u003cem\u003eAvaamo\u003c/em\u003e. [online] Available at: https://avaamo.ai/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eAvomd.com. (2024). \u003cem\u003eAI Software for Healthcare | Avo\u003c/em\u003e. [online] Available at: https://www.avomd.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eBastiongpt.com. (2025). \u003cem\u003eChatGPT for Healthcare | Medical GPT with HIPAA Compliance\u003c/em\u003e. [online] Available at: https://bastiongpt.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eBeam Health. (2025). \u003cem\u003eBeam Health | End-to-End EMR Automation for Your Practice.\u003c/em\u003e [online] Available at: https://www.beam.health/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eBeer, H. (2025). \u003cem\u003eCERSI-AI: shaping safe and effective AI in healthcare\u003c/em\u003e. [online] Hospital Healthcare Europe. Available at: https://hospitalhealthcare.com/in-depth/health-technology/cersi-ai-shaping-safe-and-effective-ai-in-healthcare/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eBell, S.K., Delbanco, T., Elmore, J.G., Fitzgerald, P.S., Fossa, A., Harcourt, K., Leveille, S.G., Payne, T.H., Stametz, R.A., Walker, J. and DesRoches, C.M. (2020). Frequency and Types of Patient-Reported Errors in Electronic Health Record Ambulatory Care Notes. \u003cem\u003eJAMA Network Open\u003c/em\u003e, [online] 3(6), p.e205867. doi:https://doi.org/10.1001/jamanetworkopen.2020.5867.\u003c/li\u003e\n\u003cli\u003eBray, I. (2025). \u003cem\u003eBMJ Group joins RADIANT to drive safe and effective digital health innovation - BMJ Group\u003c/em\u003e. [online] BMJ Group - Helping doctors make better decisions. Available at: https://bmjgroup.com/bmj-group-joins-radiant-to-drive-safe-and-effective-digital-health-innovation/ [Accessed 17 May 2025].\u003c/li\u003e\n\u003cli\u003eBrucato, C. (2023). \u003cem\u003eData Protection Policy\u003c/em\u003e. [online] Available at: https://www.gmc-uk.org/-/media/documents/data-protection-policy-gmc-v40.pdf [Accessed 16 May 2025].\u003c/li\u003e\n\u003cli\u003eCirclehealth.co. (2025). \u003cem\u003eCircle Health - AI Agents for Recovery \u0026amp; Behavioral Programs\u003c/em\u003e. [online] Available at: https://circlehealth.co/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eCombineHealth. (2025). \u003cem\u003eAI Revenue Cycle Management Software for Healthcare | CombineHealth\u003c/em\u003e. [online] Available at: https://www.combinehealth.ai/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eContinuousCare (2020). \u003cem\u003eAll-in-One Virtual Practice with AI Scribe | ContinuousCare\u003c/em\u003e. [online] ContinuousCare. Available at: https://continuouscare.io/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eCorti.ai. (2016). \u003cem\u003eCorti Assistant - AI Scribe for Clinical Documents\u003c/em\u003e. [online] Available at: https://assistant.corti.ai/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eDeepCura AI . (2025). \u003cem\u003eDeepCura AI\u003c/em\u003e. [online] Available at: https://www.deepcura.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eDeepscribe.ai. (2025). \u003cem\u003eDeepScribe AI Medical Scribe\u003c/em\u003e. [online] Available at: https://www.deepscribe.ai/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eEleos Health. (2024). \u003cem\u003eAutomation and AI Solutions for Healthcare | Eleos Health\u003c/em\u003e. [online] Available at: https://eleos.health/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eElion.health. (2025). Available at: https://elion.health/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eEmpathia.ai. (2025). \u003cem\u003eEmpathia AI - AI Medical Scribe for Clinicians\u003c/em\u003e. [online] Available at: https://www.empathia.ai/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eEngland, N. (2023). \u003cem\u003eNHS England\u0026raquo; Digital clinical safety assurance\u003c/em\u003e. [online] England.nhs.uk. Available at: https://www.england.nhs.uk/long-read/digital-clinical-safety-assurance/#:~:text=What%20are%20DCB0129%20and%20DCB0160,of%20their%20health%20IT%20software. [Accessed 15 May 2025].\u003c/li\u003e\n\u003cli\u003eEngland, N. (2025). \u003cem\u003eNHS England\u0026raquo; Guidance on the use of AI-enabled ambient scribing products in health and care settings\u003c/em\u003e. [online] England.nhs.uk. Available at: https://www.england.nhs.uk/long-read/guidance-on-the-use-of-ai-enabled-ambient-scribing-products-in-health-and-care-settings/ [Accessed 16 May 2025].\u003c/li\u003e\n\u003cli\u003eEnzo.health. (2025). \u003cem\u003eHome Health AI and Documentation Software | Enzo Health\u003c/em\u003e. [online] Available at: https://www.enzo.health/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eEpstein, J.A., Cofrancesco, J., Beach, M.C., Bertram, A., Hedian, H.F., Mixter, S., Yeh, H.-C. and Berkenblit, G. (2020). Effect of Outpatient Note Templates on Note Quality: NOTE (Notation Optimization through Template Engineering) Randomized Clinical Trial. \u003cem\u003eJournal of General Internal Medicine\u003c/em\u003e, 36(3), pp.580\u0026ndash;584. doi:https://doi.org/10.1007/s11606-020-06188-0.\u003c/li\u003e\n\u003cli\u003eEvans, K., Papinniemi, A., Ploderer, B., Nicholson, V., Hindhaugh, T., Vuvan, V., Cowley, N., Tariq, A. and Thomson, H. (2025). Impact of using an AI scribe on clinical documentation and clinician-patient interactions in allied health private practice: perspectives of clinicians and patients. \u003cem\u003eMusculoskeletal Science and Practice\u003c/em\u003e, [online] 78, p.103333. doi:https://doi.org/10.1016/j.msksp.2025.103333.\u003c/li\u003e\n\u003cli\u003eexpert.ai. (2025). \u003cem\u003eExpert.ai \u0026ndash; AI to Understand Data and Decide Better\u003c/em\u003e. [online] Available at: https://www.expert.ai/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eFazakarley, C.A., Breen, M., Leeson, P., Thompson, B. and Williamson, V. (2023). Experiences of using artificial intelligence in healthcare: a qualitative study of UK clinician and key stakeholder perspectives. \u003cem\u003eBMJ Open\u003c/em\u003e, [online] 13(12), p.e076950. doi:https://doi.org/10.1136/bmjopen-2023-076950.\u003c/li\u003e\n\u003cli\u003eFireflies.ai. (2025). \u003cem\u003eFireflies.ai | AI Teammate to Transcribe, Summarize, Analyze Meetings, Real Time AI Note Taker\u003c/em\u003e. [online] Available at: https://fireflies.ai/?utm_source=google\u0026amp;utm_medium=cpc\u0026amp;utm_campaign=brand_generic\u0026amp;utm_content=t1_eng\u0026amp;utm_term=fireflies%20ai\u0026amp;gad_source=1\u0026amp;gad_campaignid=23121070448\u0026amp;gbraid=0AAAAA-Os34pEfRBUfmRquvtPPQTlqlBT1\u0026amp;gclid=CjwKCAiAlMHIBhAcEiwAZhZBUoylRH1K2zGIn7l6gyRE7aG0Tq9QqKaIzEhdPnBRhBKps_KwbcorYRoC8wQQAvD_BwE [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eGetnudgeai.com. (2025). \u003cem\u003eThe Best AI Scribe and Coding Platform - Nudge AI\u003c/em\u003e. [online] Available at: https://getnudgeai.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eGOV.UK. (2014). \u003cem\u003eMedical devices: software applications\u003c/em\u003e. [online] Available at: https://www.gov.uk/government/publications/medical-devices-software-applications-apps [Accessed 17 May 2025].\u003c/li\u003e\n\u003cli\u003eGuidance: Medical device stand-alone software including apps (including IVDMDs) v1.10f. (n.d.). Available at: https://assets.publishing.service.gov.uk/media/64a7d22d7a4c230013bba33c/Medical_device_stand-alone_software_including_apps__including_IVDMDs_.pdf.\u003c/li\u003e\n\u003cli\u003eHardianhealth.com. (2025). \u003cem\u003eHaRi\u003c/em\u003e. [online] Available at: https://hari.hardianhealth.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eHealth (2025). \u003cem\u003eFit for the future: 10 Year Health Plan for England - executive summary (accessible version)\u003c/em\u003e. [online] GOV.UK. Available at: https://www.gov.uk/government/publications/10-year-health-plan-for-england-fit-for-the-future/fit-for-the-future-10-year-health-plan-for-england-executive-summary [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003ehealthorbit.ai. (2025). \u003cem\u003eFree AI Medical Scribe for Healthcare | HealthOrbit AI\u003c/em\u003e. [online] Available at: https://healthorbit.ai/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eHeidihealth.com. (2022). \u003cem\u003eHeidi AI\u003c/em\u003e. [online] Available at: https://www.heidihealth.com/en-gb [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eInsighthealth.ai. (2025). \u003cem\u003eInsight Health | AI Clinician Agents\u003c/em\u003e. [online] Available at: https://www.insighthealth.ai/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eJotPsych. (2024). \u003cem\u003eJotPsych\u003c/em\u003e. [online] Available at: https://www.jotpsych.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eKernberg, A., Gold, J.A. and Mohan, V. (2024). Using ChatGPT-4 to Create Structured Medical Notes From Audio Recordings of Physician-Patient Encounters: Comparative Study. \u003cem\u003eJournal of Medical Internet Research\u003c/em\u003e, [online] 26, p.e54419. doi:https://doi.org/10.2196/54419.\u003c/li\u003e\n\u003cli\u003eKiwipen.com. (2025). \u003cem\u003eCompliance | Kiwipen\u003c/em\u003e. [online] Available at: https://www.kiwipen.com/legal/compliance [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eKnowtex.ai. (2025). \u003cem\u003eKnowtex - Voice AI Automated Clinical Workflows\u003c/em\u003e. [online] Available at: https://www.knowtex.ai/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eKumpunen, S., Lobont, C., Garrard, L., Fisher, D., Lau, R., Fallica, G. and Fisher, R. (2025). \u003cem\u003eHow are GPs using AI? Insights from the front line\u003c/em\u003e. [online] Available at: https://www.nuffieldtrust.org.uk/sites/default/files/2025-12/How%20are%20GPs%20using%20AI_%20final.pdf [Accessed Dec. 2025].\u003c/li\u003e\n\u003cli\u003eKwong, M.T., Stell, D. and Akinluyi, E. (2021). Medical Device Regulation from a Health Service Provider\u0026rsquo;s Perspective. \u003cem\u003eProsthesis\u003c/em\u003e, [online] 3(3), pp.261\u0026ndash;266. doi:https://doi.org/10.3390/prosthesis3030025.\u003c/li\u003e\n\u003cli\u003eLanier, C., Dominic\u0026eacute; Dao, M., Baer, D., Haller, D.M., Sommer, J. and Junod Perron, N. (2021). How Do Patients Want Us to Use the Computer During Medical Encounters?\u0026mdash;A Discrete Choice Experiment Study. \u003cem\u003eJournal of General Internal Medicine\u003c/em\u003e, 36(7), pp.1875\u0026ndash;1882. doi:https://doi.org/10.1007/s11606-021-06753-1.\u003c/li\u003e\n\u003cli\u003eLi, B., Crampton, N., Yeates, T., Lei, Y., Tian, X. and Truong, K.N. (2021). Automating Clinical Documentation with Digital Scribes: Understanding the Impact on Physicians. \u003cem\u003eCHI\u003c/em\u003e. doi:https://doi.org/10.1145/3411764.3445172.\u003c/li\u003e\n\u003cli\u003eLyons, P.G., Rojas, J.C., Bewley, A.F., Malone, S.M. and Lekshmi Santhosh (2024). Validating the Physician Documentation Quality Instrument for Intensive Care Unit\u0026ndash;Ward Transfer Notes. \u003cem\u003eATS Scholar\u003c/em\u003e, 5(2), pp.274\u0026ndash;285. doi:https://doi.org/10.34197/ats-scholar.2023-0094oc.\u003c/li\u003e\n\u003cli\u003eLyrebirdhealth.com. (2025). \u003cem\u003ePrivacy \u0026amp; Security - Lyrebird Health\u003c/em\u003e. [online] Available at: https://www.lyrebirdhealth.com/uk/uk-compliance [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eMarianaai.com. (2024). \u003cem\u003eMarianaAI - Your Supercharged AI Medical Scribe\u003c/em\u003e. [online] Available at: https://marianaai.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eMarino, F., Alby, F., Zucchermaglio, C. and Marilena Fatigante (2023). Digital technology in medical visits: a critical review of its impact on doctor-patient communication. \u003cem\u003eFrontiers in Psychiatry\u003c/em\u003e, [online] 14. doi:https://doi.org/10.3389/fpsyt.2023.1226225.\u003c/li\u003e\n\u003cli\u003eMaticinside.ai. (2022). \u003cem\u003eHome\u003c/em\u003e. [online] Available at: https://maticinside.ai/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eMdhub.ai. (2025). \u003cem\u003eThe Clinical AI Assistant for Behavioral Health Clinicians | mdhub\u003c/em\u003e. [online] Available at: https://www.mdhub.ai/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eMEDILIT - AI Medical Scribe (The most accurate). (2023). \u003cem\u003eMEDILIT\u003c/em\u003e. [online] Available at: https://medilit.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eMedTech Europe. . IVDR \u0026amp; MDR Survey Results 2024. medtecheurope.org. (2024). Available at: https://www.gov.uk/government/publications/medical-devices-software-applications-apps [Accessed 1 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eMennella, C., Maniscalco, U., Pietro, G.D. and Esposito, M. (2024). Ethical and regulatory challenges of AI technologies in healthcare: A narrative review. \u003cem\u003eHeliyon\u003c/em\u003e, [online] 10(4), pp.e26297\u0026ndash;e26297. doi:https://doi.org/10.1016/j.heliyon.2024.e26297.\u003c/li\u003e\n\u003cli\u003eMentalyc.com. (2021). \u003cem\u003eMentalyc - AI Powered Notes \u0026amp; Insights for Deeper Therapy\u003c/em\u003e. [online] Available at: https://www.mentalyc.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eMesk\u0026oacute;, B. and Topol, E.J. (2023). The imperative for regulatory oversight of large language models (or generative AI) in healthcare. \u003cem\u003enpj Digital Medicine\u003c/em\u003e, [online] 6(1). doi:https://doi.org/10.1038/s41746-023-00873-0.\u003c/li\u003e\n\u003cli\u003eMhra.gov.uk. (2025). \u003cem\u003ePARD\u003c/em\u003e. [online] Available at: https://pard.mhra.gov.uk/manufacturer-search/Microsoft%20DAX%20 [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eMoore, N. (2025). \u003cem\u003eCompliance with NHS AI guidance and assurance standards - Lexacom\u003c/em\u003e. [online] Lexacom. Available at: https://www.lexacom.co.uk/compliance-with-nhs-ai-guidance-and-assurance-standards/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eMutuohealth.com. (2025). \u003cem\u003eAutoScribe | AI-Powered Medical Transcription and Note Generation\u003c/em\u003e. [online] Available at: https://www.mutuohealth.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eNabla.com. (2025). \u003cem\u003eNabla \u0026middot; Enjoy care again\u003c/em\u003e. [online] Available at: https://www.nabla.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eNational Data Guardian (2020). \u003cem\u003eThe Eight Caldicott Principles\u003c/em\u003e. [online] National Data Guardian, pp.1\u0026ndash;2. Available at: https://assets.publishing.service.gov.uk/media/5fcf9b92d3bf7f5d0bb8bb13/Eight_Caldicott_Principles_08.12.20.pdf.\u003c/li\u003e\n\u003cli\u003eNextGen Healthcare (2025). \u003cem\u003eNextGen Healthcare\u003c/em\u003e. [online] NextGen. Available at: https://www.nextgen.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eNguyen, A.D., White, S.J., Tse, T., Cartmill, J.A., Roger, P., Hatem, S. and Willcock, S.M. (2024). Communication during telemedicine consultations in general practice: perspectives from general practitioners and their patients. \u003cem\u003eBMC Primary Care\u003c/em\u003e, 25(1). doi:https://doi.org/10.1186/s12875-024-02576-1.\u003c/li\u003e\n\u003cli\u003eNicholson, W., Gerke, S. and Cohen, I.G. (2024). Liability for use of artificial intelligence in medicine1. \u003cem\u003eEdward Elgar Publishing eBooks\u003c/em\u003e, [online] pp.150\u0026ndash;166. doi:https://doi.org/10.4337/9781802205657.ch09.\u003c/li\u003e\n\u003cli\u003eNora AI. (2025). \u003cem\u003eNora AI\u003c/em\u003e. [online] Available at: https://www.noramed.ai/ [Accessed 16 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eNuance Communications. (2025). \u003cem\u003eNuance Communications\u003c/em\u003e. [online] Available at: https://www.nuance.com/index.html [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eOmniMD. (2025). \u003cem\u003eEHR, Billing \u0026amp; AI Solutions to Grow Your Practice - OmniMD\u003c/em\u003e. [online] Available at: https://omnimd.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eOrbdoc.com. (2025). \u003cem\u003eOrbDoc - AI Medical Scribe for Physicians\u003c/em\u003e. [online] Available at: https://orbdoc.com/index.html [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eOskrochi, Y., Roy-Highley, E., Grimes, K. and Shah, S. (2025). Digitally Unsafe? Digital health technology compliance with clinical safety standards within the NHS in England: a national cross-sectional study. (Preprint). \u003cem\u003eJournal of Medical Internet Research\u003c/em\u003e. doi:https://doi.org/10.2196/80076.\u003c/li\u003e\n\u003cli\u003ePabau. (2025). \u003cem\u003eThe all-in-one Practice Management Software you need to run your Business\u003c/em\u003e. [online] Available at: https://pabau.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003ePatientnotes.ai. (2025). \u003cem\u003ePatient Notes\u003c/em\u003e. [online] Available at: https://patientnotes.ai/us [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003ePatientnotes.app. (2025). \u003cem\u003ePatientNotes - Let AI write your clinical notes for you.\u003c/em\u003e [online] Available at: https://www.patientnotes.app/auth/signup [Accessed 28 May 2025].\u003c/li\u003e\n\u003cli\u003ePlaybackhealth.com. (2025). \u003cem\u003eAi Ambient Listening Scribe for Healthcare | Playback Health\u003c/em\u003e. [online] Available at: https://www.playbackhealth.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eRacgp.org.au. (2018). \u003cem\u003eRACGP - Artificial intelligence (AI) scribes\u003c/em\u003e. [online] Available at: https://www.racgp.org.au/running-a-practice/technology/business-technology/artificial-intelligence-ai-scribes [Accessed 17 May 2025].\u003c/li\u003e\n\u003cli\u003eReddy, S. (2024). Generative AI in healthcare: an implementation science informed translational path on application, integration and governance. \u003cem\u003eImplementation Science\u003c/em\u003e, [online] 19(1). doi:https://doi.org/10.1186/s13012-024-01357-9.\u003c/li\u003e\n\u003cli\u003eS10.ai. (2025). \u003cem\u003eBest Ambient AI Medical Scribe for any EHR | S10.AI\u003c/em\u003e. [online] Available at: https://s10.ai/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eScribe, A. (2025). \u003cem\u003eChartnote | AI Scribe, Medical Dictation, Snippets\u003c/em\u003e. [online] Chartnote.com. Available at: https://chartnote.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eScribeAmerica. (2023). \u003cem\u003eScribeAmerica Commercial - ScribeAmerica\u003c/em\u003e. [online] Available at: https://www.scribeamerica.com/video/scribeamerica-commercial/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eScribeberry (2025). \u003cem\u003eScribeberry | Never Chart Again\u003c/em\u003e. [online] Scribeberry. Available at: https://scribeberry.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eScribeEMR . (2025). \u003cem\u003eBest Virtual Medical Scribing Services | Remote Scribe Company\u003c/em\u003e. [online] Available at: https://www.scribeemr.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eScribemd.ai. (2025). \u003cem\u003eScribeMD.ai - Your AI medical assistant\u003c/em\u003e. [online] Available at: https://www.scribemd.ai/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eScribePT. (2025). \u003cem\u003eThe Best AI Scribe for Rehab Therapists l HIPAA Compliant\u003c/em\u003e. [online] Available at: https://www.scribept.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eSecinaro, S., Calandra, D., Secinaro, A., Muthurangu, V. and Biancone, P. (2021). The role of artificial intelligence in healthcare: a structured literature review. \u003cem\u003eBMC Medical Informatics and Decision Making\u003c/em\u003e, [online] 21(1). Available at: https://link.springer.com/article/10.1186/s12911-021-01488-9.\u003c/li\u003e\n\u003cli\u003eService.gov.uk. (2023). \u003cem\u003eSupplier Ambient Voice Technology (AVT) Self Certified Registry for the use of AI-Enabled Ambient Scribing Products in Health and Care Settings - Find a Tender\u003c/em\u003e. [online] Available at: https://www.find-tender.service.gov.uk/Notice/069369-2025?origin=SearchResults\u0026amp;p=138 [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eSeuren, L.M., Sakari Ilom\u0026auml;ki, Evi Dalmaijer, Shaw, S.E. and Wyke (2024). Communication in Telehealth: A State-of-the-Art Literature Review of Conversation-Analytic Research. \u003cem\u003eResearch on language and social interaction\u003c/em\u003e, 57(1), pp.73\u0026ndash;90. doi:https://doi.org/10.1080/08351813.2024.2305045.\u003c/li\u003e\n\u003cli\u003eSimboalphus.com. (2025). \u003cem\u003eSimboAlpus - Voice AI Assistant for Doctors for EMR Dictation\u003c/em\u003e. [online] Available at: https://secure.simboalphus.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eSinnott, C., Moxey, J.M., Marjanovic, S., Leach, B., Hocking, L., Ball, S., Georgiadis, A., Lam\u0026eacute;, G., Willars, J. and Dixon-Woods, M. (2021). Identifying how GPs spend their time and the obstacles they face: a mixed-methods study. \u003cem\u003eBritish Journal of General Practice\u003c/em\u003e, p.BJGP.2021.0357. doi:https://doi.org/10.3399/bjgp.2021.0357.\u003c/li\u003e\n\u003cli\u003eSOAPNoteAI (2025). \u003cem\u003eBest AI SOAP Note Generator - Free Trial | SOAPNoteAI\u003c/em\u003e. [online] SOAPNoteAI.com. Available at: https://soapnoteai.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eSollof, J. (2025). \u003cem\u003eAccurx and Tandem Health roll out AI scribing tool across the NHS\u003c/em\u003e. [online] Digital Health. Available at: https://www.digitalhealth.net/2025/04/accurx-and-tandem-health-partner-to-roll-out-ai-scribing-in-the-nhs/ [Accessed 16 May 2025].\u003c/li\u003e\n\u003cli\u003eStaff Writers (2025). \u003cem\u003eNHS drops \u0026lsquo;medical device\u0026rsquo; bomb on AI scribes in UK - Medical Republic\u003c/em\u003e. [online] Medical Republic. Available at: https://www.medicalrepublic.com.au/nhs-drops-medical-device-bomb-on-ai-scribes-in-uk/117826 [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eStokel-Walker, C. (2025). The \u0026lsquo;ambient scribe\u0026rsquo; tools listening to and summarising your doctor-patient consultations. \u003cem\u003eBMJ\u003c/em\u003e, [online] 389, p.r663. doi:https://doi.org/10.1136/bmj.r663.\u003c/li\u003e\n\u003cli\u003eSully.ai. (2025). \u003cem\u003eAI For Doctors | Terms of Service - Sully\u003c/em\u003e. [online] Available at: https://www.sully.ai/data-processing-agreement [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eSwiftynote.com. (2025). \u003cem\u003eSwiftyNote\u003c/em\u003e. [online] Available at: https://www.swiftynote.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eTierney, A.A., Gayre, G., Hoberman, B., Mattern, B., Ballesca, M., Kipnis, P., Liu, V. and Lee, K. (2024). Ambient Artificial Intelligence Scribes to Alleviate the Burden of Clinical Documentation. \u003cem\u003eNEJM Catalyst\u003c/em\u003e, [online] 5(3). doi:https://doi.org/10.1056/cat.23.0404.\u003c/li\u003e\n\u003cli\u003eTierney, A.A., Gayre, G., Hoberman, B., Mattern, B., Ballesca, M., Wilson, S.B., Castilla, K., Lau, C.S., Kipnis, P., Liu, V. and Lee, K. (2025). Ambient Artificial Intelligence Scribes: Learnings after 1 Year and over 2.5 Million Uses. \u003cem\u003eNEJM Catalyst\u003c/em\u003e, [online] 6(5). doi:https://doi.org/10.1056/cat.25.0040.\u003c/li\u003e\n\u003cli\u003eTortus. (2025). \u003cem\u003eTORTUS | Medical AI Healthcare Assistant | Doctor Approved\u003c/em\u003e. [online] Available at: https://tortus.ai/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eTrytwofold.com. (2025). \u003cem\u003eBest AI Medical Scribe for Clinicians | Twofold Health\u003c/em\u003e. [online] Available at: https://www.trytwofold.com/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eTzanou, M. (2020). \u003cem\u003eHealth Data Privacy under the GDPR\u003c/em\u003e. Routledge.\u003c/li\u003e\n\u003cli\u003eUpheal.io. (2025). \u003cem\u003eUpheal | AI progress notes\u003c/em\u003e. [online] Available at: https://www.upheal.io/ [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003cli\u003eVero Scribe (2025). \u003cem\u003eVero Scribe\u003c/em\u003e. [online] Vero Scribe. Available at: https://secure.veroscribe.com/consultation [Accessed 15 Nov. 2025].\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"University College London","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"AI medical scribes, ambient voice technology, digital health regulation, UK Medical Devices Regulations, MHRA registration, NHS governance, GDPR compliance, clinical safety standards, software as a medical device, healthcare AI transparency","lastPublishedDoi":"10.21203/rs.3.rs-8994700/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8994700/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eAI-enabled digital medical scribes are increasingly promoted as a solution to documentation burden and burnout in UK primary care. Recent NHS guidance indicates that many such tools may fall within the scope of UK Medical Devices Regulations, yet there is a lack of support for clinicians to verify regulatory status or data governance claims.\u003c/p\u003e\u003ch2\u003eObjective\u003c/h2\u003e \u003cp\u003eTo identify AI-powered health care digital scribe (ambient voice technology, AVT) products available to UK clinicians and evaluate the transparency of their regulatory, data-protection and NHS-readiness claims.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eBetween 1\u0026ndash;15 November 2025 we conducted a structured web-based review using search engines, NHS innovation and procurement portals, regulatory databases (including the MHRA Public Access Registration Database), commercial catalogues and vendor materials. Products were eligible if they were marketed as healthcare AI scribes or consultation assistants that transcribe and/or summarise clinician\u0026ndash;patient interactions. For each tool, we extracted information on features, GDPR statements, MHRA registration, claimed NHS use and access pathways.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eWe identified 118 AI-enabled health care AVT products marketed as AI scribes. Only nine (7%) explicitly claimed Class I medical device status; of these, six could be definitively located on the MHRA register under the same or a clearly linked legal entity. An explicit GDPR or data-protection statement was present for 88 products (73%), but supporting evidence (e.g. DPIAs, UK-specific safeguards) was rarely available. In terms of market accessibility, 51 tools (42%) required direct vendor contact or a demonstration, 46 (38%) showed clear UK or NHS availability, 15 (12%) were not accessible to UK users, and access status was unclear for nine (7%).\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eThere is a marked mismatch between emerging NHS expectations for AI-enabled medical scribes and the transparency of current market offerings. Clinicians may struggle to reliably verify which tools are registered medical devices, how patient data are processed, or whether NHS clinical safety standards are met. Although NHS England has since introduced a national self-certified Ambient Voice Technology (AVT) supplier registry to improve visibility of regulatory and governance compliance, this was published after completion of data collection for this study. While the registry represents an important step toward strengthening oversight, its impact will depend on supplier participation, accuracy of self-certification, and effective dissemination to clinicians and healthcare organisations.\u003c/p\u003e","manuscriptTitle":"Trust Me, I Might be a Medical Device: The Problem with AI Scribes","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-03 09:33:15","doi":"10.21203/rs.3.rs-8994700/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"61a9e286-37e7-41d8-9dc9-8b56277daf35","owner":[],"postedDate":"March 3rd, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":63694344,"name":"Medical Informatics"}],"tags":[],"updatedAt":"2026-03-03T09:33:15+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-03 09:33:15","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8994700","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8994700","identity":"rs-8994700","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}