How easily can AI chatbots spread misinformation in audiology and otolaryngology?

How easily can AI chatbots spread misinformation in audiology and otolaryngology? | medRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (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];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-P4HH5NV'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search How easily can AI chatbots spread misinformation in audiology and otolaryngology? View ORCID Profile W. Wiktor Jedrzejczak , View ORCID Profile Agata Szkiełkowska , View ORCID Profile Danuta Raj-Koziak , Elżbieta Włodarczyk , View ORCID Profile Henryk Skarżyński , View ORCID Profile Krzysztof Kochanek doi: https://doi.org/10.1101/2025.04.24.25326281 W. Wiktor Jedrzejczak 1 Institute of Physiology and Pathology of Hearing , Mochnackiego 10 Street, Warsaw, Poland 2 World Hearing Center , Mokra 17 Street, Kajetany, Poland Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for W. Wiktor Jedrzejczak For correspondence: w.wiktor.j{at}gmail.com Agata Szkiełkowska 1 Institute of Physiology and Pathology of Hearing , Mochnackiego 10 Street, Warsaw, Poland 2 World Hearing Center , Mokra 17 Street, Kajetany, Poland Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Agata Szkiełkowska Danuta Raj-Koziak 1 Institute of Physiology and Pathology of Hearing , Mochnackiego 10 Street, Warsaw, Poland 2 World Hearing Center , Mokra 17 Street, Kajetany, Poland Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Danuta Raj-Koziak Elżbieta Włodarczyk 1 Institute of Physiology and Pathology of Hearing , Mochnackiego 10 Street, Warsaw, Poland 2 World Hearing Center , Mokra 17 Street, Kajetany, Poland Find this author on Google Scholar Find this author on PubMed Search for this author on this site Henryk Skarżyński 1 Institute of Physiology and Pathology of Hearing , Mochnackiego 10 Street, Warsaw, Poland 2 World Hearing Center , Mokra 17 Street, Kajetany, Poland Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Henryk Skarżyński Krzysztof Kochanek 1 Institute of Physiology and Pathology of Hearing , Mochnackiego 10 Street, Warsaw, Poland 2 World Hearing Center , Mokra 17 Street, Kajetany, Poland Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Krzysztof Kochanek Abstract Full Text Info/History Metrics Supplementary material Data/Code Preview PDF Abstract Background Chatbots powered by large language models (LLMs) have recently emerged as prominent sources of information. However, their ability to propagate misinformation as well as information, particularly in specialized fields like audiology and otolaryngology, remains underexplored. This study aimed to evaluate the accuracy of six popular chatbots – ChatGPT, Gemini, Claude, DeepSeek, Grok, and Mistral – in response to questions framed around a range of unproven methods in audiological and otolaryngological care. Methods A set of 50 questions was developed based on common conversations between patients and clinicians. We then posed these questions to the chatbots. We tested each chatbot 10 times to account for variable responses, producing a total of 3,000 responses. The responses were compared with correct answers based on the general opinion of 11 professionals. The consistency of the responses was evaluated by Cohen’s Kappa. Results Most chatbot responses to the majority of questions were deemed accurate. Grok consistently performed best, where its answers aligned perfectly with the opinions of the experts. Deepseek exhibited the lowest accuracy, scoring 95.8%. Mistral exhibited the lowest consistency, scoring 0.96. Conclusions Although the evaluated chatbots generally avoided endorsing scientifically unsupported methods, some of the answers given could mislead and facilitate misinformation. The best performer among the group was Grok, which provided consistently accurate responses, showing it has potential for use in clinical and educational settings. Introduction The development of artificial intelligence (AI) chatbots powered by large language models (LLMs) has transformed how information is disseminated, offering users rapid and interactive responses across various domains, including healthcare, education, and science [ 1 , 2 ]. However, the reliability of these AI-generated outputs has been questioned due to the possibility that they could endorse misinformation available on the web [ 3 ]. A key concern is the phenomenon of “hallucination”, where an AI system generates a plausible but false piece of information, often accompanied by fabricated references or unsupported claims [ 4 , 5 ]. This issue has raised alarm in critical fields such as healthcare and academia, where the implications of misinformation can be profound [ 3 ]. In healthcare, studies have highlighted the risk of an AI chatbot disseminating inaccurate medical information, potentially compromising patient safety [ 6 ]. In a clinical scenario, for example, misinformation generated by a chatbot could mislead a patient, resulting in delayed or inappropriate care. Similarly, in an academic context, the use of AI-generated content in the scientific literature poses a threat to the integrity of scholarly publications, with fabricated citations and unverifiable claims becoming a significant concern [ 5 , 7 ]. Furthermore, verifying AI-generated information remains a challenge, as chatbots often fail to provide reliable or traceable sources for their outputs [ 8 ]. In audiology and otolaryngology, AI applications have garnered increasing attention in recent years [ 9 ]. Practitioners in these fields are exploring AI’s potential in areas such as screening, diagnosis, and clinical decision support [ 10 ]. Advanced AI technologies are already being integrated into hearing aids and cochlear implants, offering enhanced sound processing and personalized user experiences [ 11 – 13 ]. Despite these advances, challenges persist, particularly in ensuring the accuracy of AI outputs. Thus, in the relatively niche fields of audiology and otolaryngology, it is possible that these areas are underrepresented in large-scale training sets, increasing the likelihood that an AI system might make an inaccurate assessment or recommendation [ 5 ]. Reviews of AI applications in these fields emphasize the importance of clinicians being aware of the limitations of AI, which need to be addressed before the technology’s potential can be successfully harnessed [ 9 , 10 ]. There are a few studies which have explored how well chatbots perform when responding to controversial or nuanced medical questions [ 8 , 14 ]. However, there is a clear lack when it comes to specifically evaluating the potential for spreading misinformation in audiology and otolaryngology. When we conducted a systematic literature search using PubMed and Scopus using the keywords ‘chatbots,’ ‘misinformation,’ ‘audiology,’ and ‘otolaryngology,’ we found only 7 relevant hits, none of which addressed the problem of AI-generated misinformation. This paper aims to address this gap by investigating the reliability of six popular LLM chatbots, focusing on how well they responded to questions about scientifically unverified methods in audiology and otolaryngology. The chatbots comprised three of the most often used (ChatGPT, Gemini, and Claude [ 15 ]) and three that have recently emerged (DeepSeek, Grok, and Mistral). We compared the outputs of the chatbots, assessing their performance and looking to see how easily they could spread misinformation. Method We compiled a set of 50 questions, framed so as to represent common misinformation themes in audiology and otolaryngology ( Table 1 ). The development involved three stages. First, we collected common patient inquiries encountered over the last 12 months during clinical consultations within our otolaryngology and audiology departments. Next, a detailed review was conducted of popular media sources – health-focused websites, social media platforms, YouTube videos, and direct-to-consumer advertisements – to identify other prevalent myths and misconceptions. Finally, we asked a group of experts to review the questions and confirm that they were clinically relevant and could affect decisions about patient care. The questions were intentionally designed to elicit clear, binary yes/no responses (i.e., they were either right or wrong, leaving no room for doubt). This was our way of ensuring that the chatbots could be consistently compared. The final questions covered a broad spectrum of topics, including hearing disorders, tinnitus management, and common otolaryngological conditions encountered in patient–clinician interactions. View this table: View inline View popup Table 1. Questions and answers key. The study evaluated the following chatbots: ChatGPT (OpenAI, version o1), Gemini (Google, version 2.5 Pro), Claude (Anthropic, 3.7 Sonnet), DeepSeek (DeepSeek AI, version R1), Grok (xAI, version 2), and Mistral (Mistral AI, version Large). The entire set of questions was posed simultaneously to each chatbot using the ChatHUB system ( https://app.chathub.gg ), which allowed responses from all chatbots to be collected simultaneously. The highest available versions of each chatbot on the ChatHUB platform were used. The questions were preceded by the prompt: “Answer the following questions, providing only the question number and the answer number. Answer: 1 – Yes, 0 – No.” To account for possible variability in responses [ 16 ], each chatbot version was tested 10 times (5 trials on one day and 5 on the following). The testing was done on 26 and 27 March 2025. This approach yielded a total of 3,000 responses for analysis (6 chatbots × 50 questions × 10 trials). After each trial the data was collected into a spreadsheet and the chatbots were reset to the “new chat” state (so as to avoid previous conversations having an effect on answers). The responses were checked against an answer key developed by an expert panel consisting of 11 specialists (4 medical doctors, 4 hearing care professionals, and 3 scientists), each with over 10 years of professional experience in audiology and otolaryngology. The most frequent expert responses to each question established the set of consensus answers (which are set out in the final column of Table 1 ). All the chatbot responses are given in the supplementary material. This research did not involve human participants and was therefore exempt from Institutional Review Board approval. Statistical Analysis All analyses were performed using MATLAB software (version 2023b; MathWorks, Natick, MA). MATLAB scripts were developed with the assistance of Claude 3.7 Sonnet. Accuracy was calculated as the proportion of correct responses provided by each chatbot. Cohen’s Kappa statistic was used to evaluate inter-rater consistency. For binomially distributed data (yes/no responses), differences among chatbots were assessed using Cochran’s Q -test. Pairwise comparisons of accuracy between chatbots were conducted with McNemar’s test, while pairwise assessments of consistency utilized Fisher’s exact test. Pearson correlation was also calculated. Statistical significance was defined as p < 0.05 (95% confidence level), with Bonferroni corrections applied to adjust for multiple pairwise comparisons. Results Accuracy analysis The overall performance of each of the evaluated chatbots is summarized in Figure 1 . All chatbots achieved accuracy above 95%. Grok achieved 100% agreement with the experts, consistently providing responses that aligned with their collective opinions. DeepSeek had the lowest accuracy, 95.8%. Download figure Open in new tab Figure 1. Average percent LLM chatbot response accuracy. Error bars represent 95% confidence intervals. Horizontal lines with asterisks indicate statistically significant differences between pairs of chatbots. The Cochran Q -test revealed statistically significant differences among the chatbots ( Q = 40.89, p < 0.001). Pairwise comparisons using the McNemar test revealed several significant differences between chatbots as shown in Figure 1 . For example, Grok significantly outperformed Gemini, DeepSeek, and Mistral; there were no significant differences between ChatGPT and Claude. Consistency analysis The consistency results are shown in Figure 2 . The Cochran Q -test revealed statistically significant differences among the chatbots ( Q = 30.04, p < 0.001). Again Grok was the best with perfect consistency (1.00). Mistral showed markedly lower consistency (0.96), indicating greater variability in multiple responses to identical questions. Download figure Open in new tab Figure 2. Average chatbot response agreement (Cohen’s Kappa). Error bars represent 95% confidence intervals. Horizontal lines with asterisks indicate statistically significant differences between pairs of chatbots. Performance per question Figure 3 displays the percentage of correct responses per system for each question in the form of a heatmap. While the majority of questions were answered correctly by all systems across all trials (yielding 100% agreement), some questions revealed various degrees of inconsistency. Specifically, there were 9 questions that were sometimes answered incorrectly (3, 8, 9, 23, 33, 41, 43, 45, 47). The accuracy of responses and their consistency (Cohen’s Kappa) correlated positively ( r = 0.79, p < 0.001) indicating that accuracy and consistency in chatbot responses are not independent qualities. Download figure Open in new tab Figure 3. Heatmap showing question-specific accuracy of each LLM chatbot. Each cell represents the proportion of correct responses across 10 repetitions of the same question. Lighter colors indicate higher accuracy, while darker colors indicate lower accuracy. Discussion The present study evaluated the accuracy and consistency of six popular AI chatbots – ChatGPT, Gemini, Claude, DeepSeek, Grok, and Mistral – in addressing questions related to unproven methods in audiology and otolaryngology. In general, all AI models tested displayed high accuracy. Nevertheless, even one poor answer should be treated as a warning flag. Although the differences between the tested systems were minor, there were some notable exceptions. Grok consistently outperformed the other models, achieving 100% accuracy across all trials. ChatGPT was second best with 98.8% accuracy, while DeepSeek, though still mostly accurate (96.8%), had the lowest overall accuracy, highlighting a need for caution when utilizing such a system in a clinical or educational setting. An important consideration identified by this study was the correlation between accuracy and consistency, pointing to the fact that responses which were more consistent tended to be more accurate. This implies that efforts aimed at improving the consistency of a chatbot will also enhance its reliability. One positive aspect is that, for all tested chatbots, incorrect answers were only created occasionally, although one should keep in mind that a tendency for LLM AI systems to sometimes hallucinate suggests additional caution. Misleading answers become a serious matter if a patient then decides not to go to a specialist and tries some alternative treatment instead. When the patient eventually does go to a specialist it may be too late for effective help. The variability of responses can be explained in terms of the way that LLMs work. Because the text is generated quasi-randomly, it is inevitably variable. Such an approach might be good when creating a story or simulating a conversation, but it is bad when it relates to specific information (as in the case of medical treatments). These errors might also arise from the phenomenon of LLM hallucinations [ 4 , 7 ]. In a scientific context, hallucinations are most evident for references, when LLMs often make them up [ 5 ]. In the present study, it seems the system creators have used some way of suppressing hallucinations. In medical settings there have been numerous studies of ChatGPT, but there have not been as many for the other AI systems. Gemini and Claude are also quite well tested, but there is only a handful of studies with DeepSeek, Grok, and Mistral [ 17 , 18 ]. One study has already shown that Claude is the chatbot most resistant to attempts to use it to create disinformation [ 3 ]. Some other studies suggest that ChatGPT is the best for medical applications [ 19 – 21 ], although other studies give more mixed results [ 22 , 23 ]. In our study Grok consistently performed the best. Previous studies related to audiology and otolaryngology have not focused on the misinformation aspect, although sometimes it has been alluded to in a discussion or conclusion. In general, most studies have provided positive assessments of the use of chatbots in medicine, although a study focusing specifically on laryngeal cancer has questioned whether chatbot responses are safe or reliable enough for use by medical professionals [ 24 ]. Another study based on more general queries evaluated chatbot responses in a positive light, and did not report any risk for misinformation [ 25 ]. One study of recommendations by a chatbot about treatments for neck mucosal malignancy pointed out a number of weaknesses of the technology, but was nevertheless largely positive [ 26 ]. Uniformly, all studies have warned that any information provided by AI needs to be checked. We found that, depending on the chatbot, certain questions varied in their percentage correctness ( Figure 3 ). The majority of questions always received perfectly correct answers, but there were a few questions that proved difficult and gave varied results. Sometimes this might have been due to ambiguity (like question 43), but in others, like question 3, the answer ‘yes’ is hard to understand, since it is well documented that otoacoustic emissions cannot determine the exact number of underlying hair cells. On the other hand, one can reasonably ask how AI knows the correct answers to most questions. Where did it learn everything? Was it derived from training (most likely) or was all the relevant scientific literature sifted (unlikely) or was a specialist hired (very unlikely)? Whatever the answer, it seems rather ironic that all the knowledge was provided free to the AI system, while the user has to pay for it. Perhaps if there were regulations in place, LLM chatbots would evolve more slowly but more responsibly. At present one can only hope that a chatbot has given the correct answer; if there are no experts to hand, there is no easy way to check. As a final consideration, it seems to us that there is a need for society in general to be more educated about the way that AI powered chatbots work and understand that they inherently provide a range of answers. If that were to become common knowledge then a casual user would be unlikely to be misled. Our approach of only asking for yes/no answers might be considered as a limitation but it was done on purpose. We know that LLM chatbots tend to provide lengthy responses that are often not definite. So our solution was to force them to supply a definite yes or no answer. If we had asked open-ended questions it would have been impossible to rate them objectively – an evaluator can easily miss nuances among large amounts of text. Indeed, we experienced just this problem in our earlier studies [ 8 , 14 , 16 ] and so here we sought a different approach. In summary, we found that established chatbots like ChatGPT and Claude performed very well, reinforcing the utility of these widely used platforms. However, our findings also show that new platforms such as Grok can achieve top-rate performance. Since AI technology is evolving rapidly, the need to continuously seek out misinformation has become urgent. Steps toward mitigating risks are developing comprehensive guidelines, establishing robust oversight mechanisms, and educating people, including health professionals, about how to critically assess AI-generated information. We need a better understanding of how AI can induce misinformation so that programmers can improve the accuracy and reliability of their creations. Conclusions AI chatbots have the potential to enhance patient care in otolaryngology and audiology, but they also present challenges due to the risk of spreading misinformation. Although we found that chatbots generally avoided endorsing controversial therapies, the inherent randomness of how they generate text can sometimes lead to misinterpretation, especially for a naive patient. Given that even a single incorrect or ambiguous response may pose a health risk, robust guardrails – such as clear disclaimers, required references, and expert supervision – are necessary before these AI-powered tools can be fully integrated into clinical practice. In summary, while AI chatbots offer major benefits, their potential to spread misinformation cannot be overlooked. Concerted efforts are needed in monitoring, developing guidelines, and educating professionals to ensure that AI technologies contribute positively and do not compromise patient safety. Data Availability All data produced are attached as a supplementary file. Acknowledgment The authors would like to thank dr Andrew Bell for comments on an earlier version of this manuscript. ChatGPT (OpenAI) was used for initial brainstorming and language suggestions; all text was subsequently revised by the authors and a native English-speaking editor to ensure accuracy and clarity. References 1. ↵ Wang , D. ; Zhang , S. Large Language Models in Medical and Healthcare Fields: Applications, Advances, and Challenges . Artif. Intell. Rev . 2024 , 57 , 299 , doi: 10.1007/s10462-024-10921-0 . OpenUrl CrossRef 2. ↵ Lo , C.K. What Is the Impact of ChatGPT on Education? A Rapid Review of the Literature . Educ. Sci . 2023 , 13 , 410 , doi: 10.3390/educsci13040410 . OpenUrl CrossRef 3. ↵ Menz , B.D. ; Kuderer , N.M. ; Bacchi , S. ; Modi , N.D. ; Chin-Yee , B. ; Hu , T. ; Rickard , C. ; Haseloff , M. ; Vitry , A. ; McKinnon , R.A. ; et al. Current Safeguards, Risk Mitigation, and Transparency Measures of Large Language Models against the Generation of Health Disinformation: Repeated Cross Sectional Analysis . BMJ 2024 , 384 , e078538. doi: 10.1136/bmj-2023-078538 . OpenUrl Abstract / FREE Full Text 4. ↵ Maleki , N. ; Padmanabhan , B. ; Dutta , K. AI Hallucinations: A Misnomer Worth Clarifying . In Proceedings of the 2024 IEEE Conference on Artificial Intelligence (CAI) ; IEEE : Singapore, Singapore, June 25 2024 ; pp. 133 – 138 . 5. ↵ Jędrzejczak , W.W. ; Pastucha , M. ; Skarżyński , H. ; Kochanek , K. Comparison of ChatGPT and Gemini as Sources of References in Otorhinolaryngology 2024 , 2024.08.12.24311896. 6. ↵ Zhang , S. ; Liau , Z.Q.G. ; Tan , K.L.M. ; Chua , W.L. Evaluating the Accuracy and Relevance of ChatGPT Responses to Frequently Asked Questions Regarding Total Knee Replacement . Knee Surg. Relat. Res . 2024 , 36 , 15 , doi: 10.1186/s43019-024-00218-5 . OpenUrl CrossRef PubMed 7. ↵ Alkaissi , H. ; McFarlane , S.I. Artificial Hallucinations in ChatGPT: Implications in Scientific Writing . Cureus 2023 , doi: 10.7759/cureus.35179 . OpenUrl CrossRef 8. ↵ Jedrzejczak , W.W. ; Kochanek , K. Comparison of the Audiological Knowledge of Three Chatbots: ChatGPT, Bing Chat, and Bard . Audiol. Neurootol . 2024 , 1 – 7 , doi: 10.1159/000538983 . OpenUrl CrossRef 9. ↵ Frosolini , A. ; Franz , L. ; Caragli , V. ; Genovese , E. ; de Filippis , C. ; Marioni , G. Artificial Intelligence in Audiology: A Scoping Review of Current Applications and Future Directions . Sensors 2024 , 24 , 7126 , doi: 10.3390/s24227126 . OpenUrl CrossRef PubMed 10. ↵ Pandey , A. ; Kaur , J. ; Kaushal , D. Transforming ENT Healthcare: Advancements and Implications of Artificial Intelligence . Indian J. Otolaryngol. Head Neck Surg . 2024 , 76 , 4986 – 4996 , doi: 10.1007/s12070-024-04885-4 . OpenUrl CrossRef PubMed 11. ↵ Umashankar , A. ; Mn , A. ; C, P. APPLICATIONS OF ARTIFICIAL INTELLIGENCE IN HEARING AIDS AND AUDITORY IMPLANTS: A SHORT REVIEW . J. Hear. Sci . 2021 , 11 , 20 – 23 , doi: 10.17430/JHS.2021.11.3.2 . OpenUrl CrossRef 12. Tasnim , N.Z. ; Ni , A. ; Lobarinas , E. ; Kehtarnavaz , N. A Review of Machine Learning Approaches for the Personalization of Amplification in Hearing Aids . Sensors 2024 , 24 , 1546 , doi: 10.3390/s24051546 . OpenUrl CrossRef PubMed 13. ↵ Crowson , M.G. ; Lin , V. ; Chen , J.M. ; Chan , T.C.Y. Machine Learning and Cochlear Implantation-A Structured Review of Opportunities and Challenges . Otol. Neurotol. Off. Publ. Am. Otol. Soc. Am. Neurotol. Soc. Eur. Acad. Otol. Neurotol . 2020 , 41 , e36 – e45 , doi: 10.1097/MAO.0000000000002440 . OpenUrl CrossRef PubMed 14. ↵ Ratuszniak , A. ; Gos , E. ; Lorens , A. ; Skarzynski , P.H. ; Skarzynski , H. ; Jedrzejczak , W.W. Performance of ChatGPT in Pediatric Audiology as Rated by Students and Experts . J. Clin. Med . 2025 , 14 , 875 , doi: 10.3390/jcm14030875 . OpenUrl CrossRef PubMed 15. ↵ Moulaei , K. ; Yadegari , A. ; Baharestani , M. ; Farzanbakhsh , S. ; Sabet , B. ; Reza Afrash , M. Generative Artificial Intelligence in Healthcare: A Scoping Review on Benefits, Challenges and Applications . Int. J. Med. Inf . 2024 , 188 , 105474 , doi: 10.1016/j.ijmedinf.2024.105474 . OpenUrl CrossRef PubMed 16. ↵ Kochanek , K. ; Skarzynski , H. ; Jedrzejczak , W.W. Accuracy and Repeatability of ChatGPT Based on a Set of Multiple-Choice Questions on Objective Tests of Hearing . Cureus 2024 , doi: 10.7759/cureus.59857 . OpenUrl CrossRef 17. ↵ Marcaccini , G. ; Seth , I. ; Xie , Y. ; Susini , P. ; Pozzi , M. ; Cuomo , R. ; Rozen , W.M. Breaking Bones, Breaking Barriers: ChatGPT, DeepSeek, and Gemini in Hand Fracture Management . J. Clin. Med . 2025 , 14 , 1983 , doi: 10.3390/jcm14061983 . OpenUrl CrossRef PubMed 18. ↵ Şahin , M.F. ; Topkaç , E.C. ; Doğan , Ç. ; Şeramet , S. ; Özcan , R. ; Akgül , M. ; Yazici , C.M. Still Using Only ChatGPT? The Comparison of Five Different Artificial Intelligence Chatbots’ Answers to the Most Common Questions About Kidney Stones . J. Endourol . 2024 , 38 , 1172 – 1177 , doi: 10.1089/end.2024.0474 . OpenUrl CrossRef PubMed 19. ↵ Nwachukwu , B.U. ; Varady , N.H. ; Allen , A.A. ; Dines , J.S. ; Altchek , D.W. ; Williams , R.J. ; Kunze , K.N. Currently Available Large Language Models Do Not Provide Musculoskeletal Treatment Recommendations That Are Concordant With Evidence-Based Clinical Practice Guidelines . Arthrosc. J. Arthrosc. Relat. Surg. Off. Publ. Arthrosc. Assoc. N. Am. Int. Arthrosc. Assoc . 2024 , S0749-8063(24)00575-9 , doi: 10.1016/j.arthro.2024.07.040 . OpenUrl CrossRef 20. Meyer , M.K.R. ; Kandathil , C.K. ; Davis , S.J. ; Durairaj , K.K. ; Patel , P.N. ; Pepper , J.-P. ; Spataro , E.A. ; Most , S.P. Evaluation of Rhinoplasty Information from ChatGPT, Gemini, and Claude for Readability and Accuracy . Aesthetic Plast. Surg . 2024 , doi: 10.1007/s00266-024-04343-0 . OpenUrl CrossRef 21. ↵ Rewthamrongsris , P. ; Burapacheep , J. ; Trachoo , V. ; Porntaveetus , T. Accuracy of Large Language Models for Infective Endocarditis Prophylaxis in Dental Procedures . Int. Dent. J . 2024 , S0020-6539(24)01546-6 , doi: 10.1016/j.identj.2024.09.033 . OpenUrl CrossRef 22. ↵ Mondal , H. ; Komarraju , S. ; D, S.; Muralidharan , S. Assessing the Capability of Large Language Models in Naturopathy Consultation . Cureus 2024 , 16 , e59457 , doi: 10.7759/cureus.59457 . OpenUrl CrossRef PubMed 23. ↵ Fabijan , A. ; Zawadzka-Fabijan , A. ; Fabijan , R. ; Zakrzewski , K. ; Nowosławska , E. ; Polis , B. Assessing the Accuracy of Artificial Intelligence Models in Scoliosis Classification and Suggested Therapeutic Approaches . J. Clin. Med . 2024 , 13 , 4013 , doi: 10.3390/jcm13144013 . OpenUrl CrossRef PubMed 24. ↵ Ostrowska , M. ; Kacała , P. ; Onolememen , D. ; Vaughan-Lane , K. ; Sisily Joseph , A. ; Ostrowski , A. ; Pietruszewska , W. ; Banaszewski , J. ; Wróbel , M.J. To Trust or Not to Trust: Evaluating the Reliability and Safety of AI Responses to Laryngeal Cancer Queries . Eur. Arch. Otorhinolaryngol . 2024 , 281 , 6069 – 6081 , doi: 10.1007/s00405-024-08643-8 . OpenUrl CrossRef 25. ↵ Warrier , A. ; Singh , R. ; Haleem , A. ; Zaki , H. ; Eloy , J.A. The Comparative Diagnostic Capability of Large Language Models in Otolaryngology . The Laryngoscope 2024 , 134 , 3997 – 4002 , doi: 10.1002/lary.31434 . OpenUrl CrossRef PubMed 26. ↵ Washington , C.J. ; Abouyared , M. ; Karanth , S. ; Braithwaite , D. ; Birkeland , A. ; Silverman , D.A. ; Chen , S. The Use of Chatbots in Head and Neck Mucosal Malignancy Treatment Recommendations . Otolaryngol. Neck Surg . 2024 , 171 , 1062 – 1068 , doi: 10.1002/ohn.818 . OpenUrl CrossRef PubMed View the discussion thread. Back to top Previous Next Posted April 25, 2025. Download PDF Supplementary Material Data/Code Email Thank you for your interest in spreading the word about medRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following How easily can AI chatbots spread misinformation in audiology and otolaryngology? Message Subject (Your Name) has forwarded a page to you from medRxiv Message Body (Your Name) thought you would like to see this page from the medRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share How easily can AI chatbots spread misinformation in audiology and otolaryngology? W. Wiktor Jedrzejczak , Agata Szkiełkowska , Danuta Raj-Koziak , Elżbieta Włodarczyk , Henryk Skarżyński , Krzysztof Kochanek medRxiv 2025.04.24.25326281; doi: https://doi.org/10.1101/2025.04.24.25326281 Share This Article: Copy Citation Tools How easily can AI chatbots spread misinformation in audiology and otolaryngology? W. Wiktor Jedrzejczak , Agata Szkiełkowska , Danuta Raj-Koziak , Elżbieta Włodarczyk , Henryk Skarżyński , Krzysztof Kochanek medRxiv 2025.04.24.25326281; doi: https://doi.org/10.1101/2025.04.24.25326281 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Otolaryngology Subject Areas All Articles Addiction Medicine (568) Allergy and Immunology (863) Anesthesia (300) Cardiovascular Medicine (4435) Dentistry and Oral Medicine (444) Dermatology (382) Emergency Medicine (608) Endocrinology (including Diabetes Mellitus and Metabolic Disease) (1509) Epidemiology (15229) Forensic Medicine (30) Gastroenterology (1124) Genetic and Genomic Medicine (6600) Geriatric Medicine (668) Health Economics (997) Health Informatics (4536) Health Policy (1368) Health Systems and Quality Improvement (1613) Hematology (541) HIV/AIDS (1264) Infectious Diseases (except HIV/AIDS) (15916) Intensive Care and Critical Care Medicine (1103) Medical Education (623) Medical Ethics (146) Nephrology (667) Neurology (6599) Nursing (346) Nutrition (998) Obstetrics and Gynecology (1144) Occupational and Environmental Health (957) Oncology (3332) Ophthalmology (974) Orthopedics (369) Otolaryngology (420) Pain Medicine (436) Palliative Medicine (130) Pathology (663) Pediatrics (1693) Pharmacology and Therapeutics (691) Primary Care Research (711) Psychiatry and Clinical Psychology (5447) Public and Global Health (9232) Radiology and Imaging (2198) Rehabilitation Medicine and Physical Therapy (1370) Respiratory Medicine (1196) Rheumatology (593) Sexual and Reproductive Health (712) Sports Medicine (530) Surgery (712) Toxicology (99) Transplantation (289) Urology (265) (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'a0085653bde409d6',t:'MTc3OTU4NDE1MQ=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();