Using AI to Create Cultural Content: Ethnographic Insights into Everyday Ethics

Using AI to Create Cultural Content: Ethnographic Insights into Everyday Ethics | 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 Article Using AI to Create Cultural Content: Ethnographic Insights into Everyday Ethics Bronwyn Isaacs This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6926166/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 11 You are reading this latest preprint version Abstract This paper explores how small-scale content creators in New Zealand navigate the ethics of using artificial intelligence (AI) in their everyday work. Ethnographic methods including "deep hanging out" and qualitative interviews, were used to study content creators. The paper makes a number of findings. First, users of AI technology may approach it with moral and ethical intentionality, rather than as passive consumers. Second, there is tension between the pressure to create content for the "attention economy" and to maintain ethical practices, especially when dealing with culturally sensitive material. Third, small-scale, relationship-based content creation is valued for maintaining social connections and respecting cultural contexts. This paper highlights the importance of considering scale and everyday user experience when considering the ethics of generative AI. Social science/Anthropology/Social anthropology Social science/Science technology and society Figures Figure 1 Figure 2 Introduction The widespread use of generative AI is quickly changing how many people think about culture and creativity. Considerable academic and public attention is directed towards various concerns, such as the distinctions between human and machine intelligence, the enhancement of productivity alongside the loss of human jobs, the manipulation of individuals through deepfakes, and the emergence of new forms of human emotional entanglement with machines. All these issues relate to culture, that is, how humans construct social worlds and webs of meaning. This paper focuses on a particular cultural issue concerning AI: what is assumed and what is lost when culture is treated as something that can be endlessly scaled up, accessed, and shared without consultation or connection with the communities with whom that culture originates. New AI tools are being trained on datasets that scrape information from increasingly diverse sources. The implementation of new technologies, such as Generative Pretrained Transformers (GPT) that can both learn and fine-tune a large language model, Convolutional Neural Networks (CNNs) for image classification, Recurrent Neural Networks (RNNs) to handle complex patterns, and Generative Adversarial Networks (GANs) that produce fake images that appear real (Bengesi et al. 2024). The practice of using these technologies to work with increasingly larger data sets in evidence by all leading generative AI firms has not gone unchallenged (see Khattak, Cohen and Taylor 2024 for a list of copyright lawsuits). Moreover, the operational scales at which generative AI is created and operated also deserve deeper, culturally specific forms of ethical reflection. This study focuses on the work of small-scale video content creators in Waikato, New Zealand. The paper argues that even small-scale content creators are wrestling with deep ethical frameworks when it comes to the use of generative AI. These frameworks are embedded in localised social relationships and show how using AI to create culturally relevant content can be a deeply ethical practice. That generative AI research can also further ongoing colonial legacies of extraction and inequality is an ethical concern. The creation of new generative AI products has been shown to be primarily motivated by corporate financial objectives, thus shaping biases in the data collected and the kinds of racial and ethnic biases that large language models (LLMs) display (Adams 2025; Klein & Ignazio 2024, Gebru 2020). Bender et al. (2021, 616) highlight that the risks of operating at the scale of generative AI’s large language models (LLM) include bias, increasing power imbalances and giving increased authority to existing hegemonic viewpoints (Bender et al 2021). Widder, Whittacker, and West (2024, 82) also highlight the power imbalances perpetuated by such models, noting that “the present pattern in AI development takes a bigger-is-better approach when it comes to data, but. . . information on the datasets for models has become increasingly opaque, for closed and ostensibly open models alike”. This pattern of bigger is better data scraping moreover fails to do justice to the claims of data sovereignty, story sovereignty and cultural IP called for by indigenous and minority communities in different parts of the world (King, Gybele & Anderson 2015; Kukutai & Taylor 2016). Generative AI changes quickly and ethnographic projects take a long time to complete. In this paper, I meet this challenge by reporting on an ongoing ethnographic project in which I am still actively engaged, with content creators and people working in the New Zealand film industry. AI has been embraced in the international film industry including in New Zealand where it is used for handling more repetitive and “boring” tasks such as background CGI work and as a tool for reflection and refining ideas (Narayan et al 2022). There are however considerable concerns and forms of resistance to AI creative technologies expressed by people working across the film production process in many parts of the world particularly regarding labour rights, copyright and IP infringement and loss of earnings (Khattak, Cohen and Taylor 2025). Using anthropologists’ methodological attention to events and people in detail, in this paper I focus on a small number of content creators and their creative processes over a limited time period. I ask, how do independent media creators, working with limited resources engage with AI and think about AI in their work? Do they see their work as something which that is applicable for AI models and can be “scaled up”? And what philosophies and ethics are in play when they use or don’t use AI in their everyday creative activities? Ethics is understood here as the way that people as social beings wrestle with their moral obligations to others, including moments of moral breakdown and questioning (Brodwin 2013; Kleinman 2006). This presumes that ethics are not merely matters of individual reflection but can be expressed in concert with the environment and the collective in which one lives (Jackson 2011). The results of the research project discussed here indicate that people using generative AI may do so in a deeply ethical way. They may think carefully and critically about their use of AI and the risks that emerging AI technologies pose for issues of cultural respect, social relationships, and exploitative treatment of others via the use of AI technologies. Methods The two main research methodologies on which this research project is based are ethnographic “deep hanging out” and qualitative interviews. Studies of AI ethics are often treated as “lower status” that studies in AI engineering because of the emphasis on subjective forms of knowledge (Widder 2024). Studies that consider small, scale, experiential forms of knowledge can however bring to light the views of those who are not represented in larger statistical or large scale research methodologies. This style of small scale research also meets an important call by feminist and minority researches for AI research to focuses on real people and their real experiences (Zong and Matias 2024 ; Bassett 2023 ). Klein and D’Ignazio (2024, 106) for example state, “Put simply: the field of AI needs to develop more participatory, more responsible, and more humble methods for being in dialogue with impacted communities.” This paper emerges from a project that has been in process for six months, and includes at this point, 14 interviews as well as ethnographic observation, action or participatory based research occurring over a period in of about three months. The interviewees thus far come from a range of sectors in the film industry in New Zealand, including those working in areas such as film directing, copyright, and diversity and inclusion. This study has not engaged any of the generative AI companies used or discussed by the participants. However there is research value in attending to how users use generative AI even if it is in ways not intended. As Matias ( 2023 , 251) notes, “Companies have strong incentives to resist studies that could suggest they are responsible for harms involving their products. Therefore, future studies will depend on innovations in adversarial data access, ethics, and privacy.” The ethnographic research discussed in this paper thus focuses on small local content creators, working with small budgets and relying on self production and community support. The paper focuses on one particular day of filming which serves to centre the themes of ethics as they are experienced on the small-scale, local level. Ethical approval for this project was provided by the University of Waikato Human Research Ethics Committee, project FS2024-68. Ethnography’s attention the small scale, has own unique advantages to studying culture. What ethnography loses in broad applicability, Herzfeld ( 2001 :23) writes, “it gains in the sheer intensity of the ethnographic encounter- as intimacy, as privileged access, as listen to the voices silenced on the outside by those who wield great power”. Parreñas ( 2023 ) compares the work of ethnographers to moles as opposed to mining corporations. While the latter extract valuable content on a profitable and extractive scale, ethnographers who are moved by what they witness on a sensorial level are more like “tiny moles that make new connections by digging channels in the dark” (Parreñas 2023 , 258). In regards to the study of AI, van Voorst and Ahlin ( 2024 ) highlight the ability of ethnography to probe things associated with our use of technologies that may be unseen or “absent-present”, things important in people’s daily experience of AI but not assumed by those who make it. Bringing these perspectives in view, this paper presumes that there is research value in focusing on a single day of filming including the experiences views of just two content creators, as is discussed in the following sections. Results Content Creators Consider AI images Ethically At five AM on an Autumn morning in Huntly, a semi-rural area of Waikato in New Zealand, I crunched my feet down a long gravel driveway, looking to meet three local men. These men were all working in various ways in the north island of New Zealand film, video, and content creation. These men were firstly, Joe Wilson, a local content creator and social organiser, secondly, a man working as Joe’s director of photography (DP), and thirdly, Sam Troth, a content creator and activist. Sam was currently part way through a journey of walking the length of New Zealand’s north island, a journey he saw as both self-healing from his own experiences of abuse, and a way to raise awareness of abuse issues in New Zealand. Still tucked in sleeping bags and blankets at the back of the club were Sam’s two current team members, who were also participating in the walk and assisting him with his mission. Together, the three walkers took turns walking each day, covering a route stretching from the top to the bottom of New Zealand’s North Island. Sam’s plan for this walk, which he and referred to as a hikoi (a Māori word for walk or journey, often associated with a protest march) was to finish in Wellington, the country’s capital, where he and the two others walking with him would present a petition at parliament house calling for longer sentences for those convicted of sexual abuse. Joe, a social organiser and content creator from the Waikato, was there in his capacity as the producer and director of Lots of Little Fires , a production company that in Joe’s words aims “to bring to light invisible labour” of people doing local charity and community development. At a previous event, Joe explained his work to me as a bridge between community workers needing support and those who can support them, particularly with funding. When I asked Joe why he makes videos, he explained, that based on prior experience of working in youth and council work, “if I write a report, nobody will read it”. Thus, despite having no background in film work, Joe had turned his advocacy approach to working with video because he viewed it as a necessary tool. Many of the videos Joe made were for local charities and community groups operating in distinct cultural contexts. This included Manaaki Rangatahi ki Waikato , a charity providing housing for homeless Māori youth; Rainbow Hub , a community group advocating for rainbow communities; and Poutama Rites of Passage , an Indigenous community-led youth mentorship. Many of the videos that Joe maked through Lots of Little Fires thus engaged with content considered relevant to culture, including Indigenous culture. On the day in question, Joe, his DP and Sam met in Huntly, and Joe planned to spend the day getting footage and audio he could use to make short video that would help to highlight Sam’s advocacy and his contribution to his extended communities. That morning in Huntly, the night was inky black but warm under cloud cover, and we gathered outside the coast by the quiet coast motorcycle club’s concrete building. The DP took B-roll footage of Sam making a cup of coffee and checking messages on his phone while the three men (DP, Sam and Joe) joked about the amount of labour each felt compelled to do on their social media accounts. Joe explained his kaupapa was to bring the stories of communities are charities to light but that hated the need to constantly post and monitor messages online. Kaupapa is a word from the indigenous te reo Māori language meaning purpose or vision. Stemming from Māori te ao (philosophy), a kaupapa is never only for the individual. Rather the term implies social good. The term is widely used in New Zealand outside of the Maori community and allows speakers to bring a sense of ethical weight or grounding to their speech. Sam replied to Joe that his kaupapa was to spread light on sexual abuse and survival. He expanded on how this was his second walk across Aotearoa to raise awareness of these issues. Sam discussed how spending hours walking process slowed life down and provided a new perspective, “almost spiritual”. This special experience also had to be filmed. Figure 1 shows a photo taken of Sam during his morning from the perspective of the van that was following him with the Little Fires film crew inside. Joe, the DP and myself spent several hours in a van, trailing beside Sam as he walked on the side of the road. Much of the time, we chose to wait before filming because we did not want footage of Sam using his phone. Sam spent quite a bit of time on his phone, calling people and creating content so that he would have something to choose to post later. Joe’s initiative in creating promotional videos rather than writing reports, and Sam’s felt need to constantly post reflected the challenges of generating social change within an “attention economy”. By “attention economy”, scholars refer to the technological-social changes in human cognition and reliance on owing to the everyday integration of smart devices and digital communication (Paasonen, 2021 ). The attention economy as based on profit-motivated algorithms of social media platforms and the physical design of smart devices designed distract and attach users to digital interfaces, affects individuals’ concentration and ability to provide sustained attention (Kozyreva et al. 2020 ). Joe operated his production company on the belief that those working for charities and local councils that give funding to community organisations in New Zealand, either do not read, or are not swayed or containing information when provided in written form, but require aesthetically appealing and emotionally resonant forms of visual media for a message to break through. Sam and Joe both discussed the need to post and use social media accounts when they would rather be doing other things. That morning, while drinking his pre-dawn coffee, Sam continued to discuss his need to post frequent content: It’s a lot man”, and “You’ve got to know which is the content that’s right to put up”. Going on, Sam explained, “So this week I got a call asking if I could talk to this guy who was suicidal. Talk to him, have the chat; he has been through terrible things. However, you cannot post about that. That’s not right. We could put up some pretty extreme content-stuff that would get lots of views, but we gotta do what’s right. Its like. .. Here , Sam stood up walked towards the DP, his arm outstretched, “ Like - here mate here’s 20 -- like waving 20 [dollar note] at a homeless man. That is what it would be like, taking advantage. There are all the things you cannot show.” Sam then turned his phone around, providing us with a good view of his screen which shone like a small orb against the blackness of the night around us. “There are all these things you cannot show. You must be careful when making posts depicting children in danger. I tried to make an image about a child scared in their bedroom, and AI made the child’s face a cabbage patch doll. That’s not right”. Sam continued to scroll, turning his phone between his face and ours, showing us images made by ChatGPT and other apps. Some images contained dark themes related to Sam’s work in raising awareness of abuse. There were images made in the style of black and white illustrations that depicted children hiding in their bedcovers while a shadowy figure lurked at their bedroom door. Another use of generative AI, more light-hearted in nature, was an image Sam had made in the style of a viral “doll” trend, featuring an image of a plastic-looking doll packaged as if for sale in a box. The doll resembled Sam and came with a high-vis vest, such as the one Sam wore for safety on the road, a torch and a miniature toy phone, reflective of Sam’s self-perception of his social media responsibilities while out of the road. Sam’s frustration in attempting to navigate chatGPT to produce images relevant to his kaupapa is illustrative of the elisions and biases built into the LLMs on which generative AI are made. The impact of this can be considered as a form of what Kane (2014) calls the “post-optical” quality of digital interfaces and images. Using the term “post-optic”, Kane emphasises that digital images reflect the limitations of machine software more than they do the decisions of human design. Kane argued that when users create images with digital tools, they know very little about the eliminations and standardisations through which the code of images is produced. Thus, users do not determine what is seen online; rather, computer code is largely determinative of the images and thus of image culture. Kane’s argument remains extremely relevant when it comes to AI-generated images. When using generative AI to create images, the original datasets on which generative AI is created are no longer referenced directly but pertain only to a “set of distinct statistical probabilities” (Menotti 2025 , 2). The images Sam produced of scared children therefore pertained to no individual instance in which an artist considered themes of children and fear, but a statistical formula, the parameters of which were standardised and invisible. The day of filming continued and was full of creative improvisation and hustle. The theme of the day’s filming from my perspective as an outsider seemed to be getting things done with minimal resources, and making content with what was available. For example, Joe had a few ideas of locations that day where we could film. However, on arriving at one of the parks Joe had planned on, we were unsuccessful in finding a quiet spot to film. Our filming date was during school holidays, and the park was busy with foot traffic. Joe, the DP and I discussed among ourselves possibly alternatives such as filming in various people’s back gardens and other parks in town. Eventually myself, Joe and the DP drove around the rural area of Ngahinapouri, in order to find a location quiet and with a visually striking backdrop in order to film the interview with Sam, finally settling on the location in Fig. 2 . For Sam, in the terms he described and used, the use of generative AI was a tool in his everyday social media practices of posting content to social media. These posts were made, as Sam explained, to raise awareness and connect with others on a similar journey of healing. While this work was important to him, Sam felt exhausted by it, demonstrating a “posting fatigue” that is not uncommon in attention economies (Bright, Klesier & Grau 2015). Using generative AI to create images for posts was not necessarily core to Sam’s kaupapa , but as it was a way, when his resources were limited, of creating content, a task he felt to be urgent and necessary. Later that day, Joe interviewed Sam about why he spent so much time and energy raising awareness about sexual abuse. Sam expanded on his kaupapa , explaining he was doing healing work on himself and doing something good for the world and his kids. Joe turned to me and to his DP several times throughout the interview asking, “Is this getting the story arc? Is there a question I should be asking?”. Here, a human, face-to-face discussion was valuable owing the gravity of what was being discussed and the open-endedness of the creative process. Throughout the day of filming, we discussed our connection to and the necessary labour for creating content on social media. In this moment, however, the kaupapa of both Joe and Sam as content creators came through in their deeply social commitments. Content Creators Care about Creativity Within a few weeks I followed up with Joe to discuss his editing process for the footage and audio that Joe and the DP collected on the day of the shoot with Sam. Perhaps surprisingly, Joe explained to me that the goal with videos was not maximum reach or creating viral videos, but instead his goal in posting the videos was “not how many people saw the video but if the right person did”. Explaining the importance of making connections with people, Joe said emphasised filming days were typically short (usually, as with the film about Sam, only one day of filming), but that they were done in the context of ongoing relationships: “For me, it’s so important to do those earlier visits so that people get used to you, but it can be really challenging for many people to be comfortable on camera. I have all these questions written down I mean to ask them but I end up chatting about all kinds of other things.” Joe viewed the power of what he did in communicating with others through the lens of storytelling. “I come from the North of England you see”, Joe explained to me, Where I come from, we have a tradition of folk stories, of telling stories that speak to people. Here, Māori stories also have power. And as a songwriter, I see the power of story vulnerability. When you are busking, you see people respond to those songs where you share a piece of yourself. These stories explained Joe, are communicated most powerfully in small-scale settings. That’s why I call my program Lots of Little Fires , because if you scale up you lose the campfire - you lose social connections, the fire needs to be done on a smaller scale. A big fire is too hot, and a big social project likewise results in lost relationships and innovation. I worked with the youth for the local council. I was very frustrated by the lack of care they had for people. But if you get people one or one, or in a small group, then they see each other as people, they can hear them. Based on this philosophy of social connection, Joe carefully manages who has access to his filming process. In choosing a videographer to work with, Joe also emphasised that he did not look for a particular skill set or aesthetic style, but rather that the videographer needed to “be humble” because “the quality is only as good as the work and the people”. In regard to my presence at events, Joe judged, for example, that my presence at a meal-sharing service that he planned to film may lead some people to feel uncomfortable. At a free meal service, the costs of participants feeling fear of judgement and possible recognition from outsiders were high. However in regards to the filming activity discussed in this paper, that is, the filming of Sam Troth, who was himself content creator and keen to get more attention for his channel, Joe checked and learned that Sam would happy for me to follow the filming process. This focus on personal and social relationships as key to negotiating the ethics of content creation was also evident in Joe’s selective use of AI tools in his work. On the one hand, Joe quickly accepted his videographer’s recommendation of using AI tool REV to take audio files of filming content and transcribe them. When I asked Joe about this he said that the DP sent him the interviews as a REV file and this “saves a huge amount of time, work that we would have to do”. For Joe, however, the use of generative AI demands careful, ethical reflection. Speaking of his videographer and his advice regarding AI, Joe commented, “[The DP] has said that we could put the interviews into ChatGPT and get ChatGPT to select the interview content along with the timestamp and that could be really useful but I do not want to go down that road. Because the work we do is really relational and I do not want to give to chat GPT and let it die.” When I asked what he meant by “let it die”, Joe clarified, his concern was twofold, “It’s ok to use AI to save time but I don't want artificial intelligence doing the creative work that humans should be doing. Artificial intelligence should not be doing the creative work of humans. And what the people I work with give me is a gift. It’s their stories, their suffering and my job to honour that. Its not that we shouldn't use AI, but we must use it in the right way. For me, it’s the story teller tradition - connecting with people first and foremost”. Joe’s concerns with AI thus fell into two categories, protecting the category of human creativity and secondly, honouring the sharing of stories as a gift – something outside of simple transactional exchange. Joe’s concerns about human creativity as distinct from machine-generated content are reflected in wider sectors of content and film industry (Lee 2022 ). Khattak, Cohen, and Taylor ( 2025 , 2) note that the issue of creativity as a human-led process is particularly fraught in the film industry, where “artists, screenwriters, and filmmakers have increasingly pushed back, warning that AI threatens not only their jobs but also the core of artistic authorship and cultural integrity”. Joe’s unwillingness to give data to ChatGPT is also an example of what some scholars have identified as the practice of data refusal. Zong and Mattias (2024) for example, who draw on the work of Ruha Benjamin’s ( 2019 ) work on the relationship between technology and racism, defining data refusal as a practice on practice used by those whose lives may seem peripheral to the centre of tech decision making but who are directly impacted by its design. Refusing to engage in acts of data collection can be for those on the periphery of tech worlds, an act of “contest[ing] the power of data in their lives, while also making the power of data collectors more legible” (Zong and Mattias 2024 3). Joe’s use of the term “gift” moreover reflects a deeply moral position on what is distinctly human. Gifts are a key practice in human culture. Anthropologists have shown gift giving to be possibly universal and certainly an ancient and enduring form of social obligation, binding the giver to the receiver in a debt of obligation (Graeber, 2011 ). Gifts can also frequently have a sacred quality, as they can entail giving a part of oneself to another (Mauss 1925). By linking the gift of a story to a wider tradition of storytelling, Joe’s ethical perspective also reflects a broader intentional ethical and epistemological view of knowledge sharing and storytelling. In this epistemology, as practiced in many Indigenous and other minority communities, as eloquently outlined by scholars such as Frederick (2003), Wilson ( 2020 ), Teaiwa ( 2024 ), and Funaki (2025), knowledge shared in the context of community and social relationships holds a strong ethical quality that binds social ties and can increase capacity for shared responsibility and collective action. Curious as to whether Joe’s concerns regarding AI, the safeguarding of the “gift” of shared stories reflected more broadly with the small communities and charities he works with, I asked Joe, “Do the charities and communities share your perspective on AI or do not really care?”. Joe replied, “Its not that they don't care. It’s that they may choose to use generative AI when they need an output and don't have many resources. They don't have the support and resources they need, so they may choose to rely on it”. Contesting or refusing data extraction is therefore an example of Joe expressing his kaupapa , his ethical commitment, and purpose. As Joe eloquently explained, the metaphor of a little fire is that a fire that is too large is too hot. Here it is not size per see that is the issue, but rather the dangers of lost social connections and respectful engagement of others around the metaphorical fire because there is value in paying attention to how social connections alter at different scales. To be clear, the issue is not with scale per see, but in expansion without reference to its social and cultural impact. This point is forcefully made by Tsing ( 2015 ), who argues that many capitalist projects from plantation slavery to computing attempt to reduce complex and diverse information about people and the environment to comparable units. A computing-based scalable unit such as a pixel is designed to be endlessly replicable, allowing for a potentially endless process of zooming, vision going in, out, and across. Such scalability explains that Tsing ( 2015 , 38) “banishes meaningful diversity”. Computing based attempts to treat things as uniform without paying attention to local questions of culture or ethics risk replicating colonial models of expansion and extraction and inform the most extractive versions of contemporary or late capitalism. In the context of the work of small content creators Joe and Sam, generative AI was difficult to ignore owing to the opportunity it offered to decrease labour time. As content creators, Sam and Joe were not operating according to traditional capitalist models in that they were not seeking commercial funding or payments for their products. Nonetheless, the broader patterning of the attention economy has impacted the media industry in which their work was created. Even within the attention economy, however, the creative process, including that of AI use, was navigated by Sam and Joe through a committed ethical framework. For Joe, it was about the integrity of the small community, where people know and understand each other. For Sam, a larger network was sought, but it had to be handled on a person-to-person basis. For both, it was important to allow for individual suffering and circumstances to shape the ethics of how content is created and treated. Conclusion In attending to the small scale, this paper focused on ethical and epistemological arguments for better understanding both the limits and harms of AI in regard to representation, study, and creative engagement with culture. Content creators such as Sam and Joe view AI as a tool to save time on mundane tasks but are cautious about using it for creative work or handling sensitive stories. This research project should not be assumed to have broad statistical replicability. Nonetheless, in paying attention to what is important to local content creators, this research demonstrates the importance of considering ethical concerns as part of users’ experiences. This research project provides evidence of content creators balancing the efficiency of generative AI with ethical concerns pertaining to preserving human creativity and respecting cultural content. The concept of "kaupapa" (purpose/vision) is important to these content creators' ethical decisions regarding AI use. According to the “bigger is better” models of AI design, there are few limits to the kinds of media and content which future datasets might access. There are however possible alternatives to this presumed model of continued extractive data scraping. Legally, it is appropriate to consider increased protection for individual privacy, copyright, and IP holders, in addition to cultural sovereignty over community cultural content. For tech companies engaging in generative AI data training, there is also the possibility of “algorithm abandonment” recognising the social harms of existing algorithms and when needed, agreeing to dispense with them (Johnson et al. 2024 ). Primarily, this paper calls for researchers in academia and technology to better appreciate the way in which AI users use AI with moral and ethical intentionality. AI users are not replicable standardised units. They are complex and diverse owing to their culture, their ethics and commitments to their communities. Declarations Data Availability The data was obtained through informal interviews and participant observation. It cannot be shared in accordance with the ethical procedures in which this study was conducted, including respect for participants and the sensitivity of their personal experiences. While the main participants mentioned in the ethnography chose not to be anonymous, they did not give permission for all interview data and research notes to be shared. Ethics: Ethical approval for this project was provided by the University of Waikato Human Research Ethics Committee, project FS2024-68 on 16 December 2024. The procedures used in this study adhered to the principles of the Declaration of Helsinki. Informed Consent: Informed consent to participate in the study was obtained from all participants on 18 March 2025 and 15 April 2025 respectively. Participants received an information sheet outlining the nature and aims of the research and were given the opportunity to ask questions prior to providing consent. Consent was documented via signed consent forms; in cases where participants preferred, verbal consent was obtained and recorded by the researcher. All participants were offered the option to remain anonymous. The individuals discussed directly in this article are content creators engaged in advocacy work and have chosen to be identified. Funding: Funding was provided by the University of Waikato. Funding was provided directly to the author’s individual research trust account. This funding was provided for research purposes. Beyond this there is no particular name or grant associated with the funding. Author Contribution B.I. did the research and wrote the manuscript. References Adams R, Adeleke F, Junck L, Alayande A, Segun S, Gupta A, Zaman M et al (2025) Mapping the Potentials and Limitations of Using Generative AI Technologies to Address Socio-Economic Challenges in LMICs. VeriXiv 2:57. https://doi.org/10.12688/verixiv.948.1 Bassett C (2023) The Cruel Optimism of Technological Dreams. In: Jude Browne (ed.), Stephen Cave (ed.), Eleanor Drage (ed.), Kerry McInerney (ed.) Feminist AI: Critical Perspectives on Algorithms, Data, and Intelligent Machines, p 258–273 Bender EM, Gebru T, McMillan-Major A, Shmitchell S On the dangers of stochastic parrots: Can language models be too big?列. In: Proceedings of the 2021 ACM conference on fairness, accountability, and transparency, pp. 610–623 Bengesi S, El-Sayed H, Sarker MK, Houkpati Y, Irungu J, Oladunni T (2024) Advancements in Generative AI: A Comprehensive Review of GANs, GPT, Autoencoders, Diffusion Model, and Transformers. IEEe Access. 69813–69837 Benjamin R (2019) Captivating technology: Race, carceral technoscience, and liberatory imagination in everyday life. Duke University Press, Durham Bright LF, Kleiser SB, Grau SL (2015) Too much Facebook? An exploratory examination of social media fatigue. Comput Hum Behav 44:148–155 Brodwin P (2013) Everyday ethics: Voices from the front line of community psychiatry. Univ of California, Berkeley Frederick Marla (2003) Between Sundays: Black women and everyday struggles of faith. Univ of California, Berkeley Funaki H, Macdonald L, Knox J, McKinnon D (2025) Living in the Telling: Indigenous storytelling of post covid desires for academia. Art Res Int Transdisciplinary J 8(2):499–518 Gebru T (2020) Race and gender. The Oxford handbook of ethics of AI, 4, 253 Graeber D (2011) Debt: The First 5,000 Years. Melville House, London Herzfeld M (2001) Anthropology: Theoretical practice in culture and society. Malden, Blackwell Publishers Jackson M (2011) Life within limits: Well-being in a world of want. Duke University Press, Durham Johnson N, Moharana S, Harrington C, Andalibi N, Heidari H, Eslami M (2024) The Fall of an algorithm: Characterizing the dynamics towards abandonment. In Proceedings of the 2024 ACM Conference on Fairness, Accountability, and Transparency, p 337–358 Khattak M, Cohen SE, Taylor K (2025) Who Holds the Camera? Filmmaking Justice in the Era of Generative AI. GRACE: Global Review of AI Community Ethics, 3(1) King L, Gubele R, Anderson JR (2015) Survivance, Sovereignty, and Story. Denver, University Press of Colorado Klein L, D'Ignazio C (2024) Data feminism for AI. In Proceedings of the 2024 ACM Conference on Fairness, Accountability, and Transparency, June 2024, p100-112 Kleinman A (2006) What really matters: Living a moral life amidst uncertainty and danger. Oxford University Press, Oxford Kozyreva A, Lewandowsky S, Hertwig R (2020) Citizens versus the internet: Confronting digital challenges with cognitive tools. Psychol Sci Public Interest 21(3):103–156 Kukutai T, Taylor J (2016) Indigenous data sovereignty: Towards an agenda. ANU, Canberra Lee HK (2022) Rethinking creativity: creative industries, AI and everyday creativity. Media Cult Soc 44(3):601–612 Matias JN (2023) Humans and algorithms work together—so study them together. Nature 617(7960):248–251 Mauss M (2024) [1925]). The Gift: the form and reason for exchange in archaic societies Oxon. Routledge Menotti G (2025) The model is the museum: Generative AI and the expropriation of cultural heritage. AI & SOCIETY, pp 1–5 Narayan AD, Caillard D, Matthews J, Nairn A (2022) Artificial imagination: Industry attitudes on the impact of AI on the visual effects process. Interactions: Stud Communication Cult 13(2):113–131 Paasonen S (2021) Dependent, distracted, bored: Affective formations in networked media. MIT Press, Cambridge Parreñas JS (2023) Ethnography after anthropology: Become moles, not mining corporations. Am Ethnologist 50(3):453–461 Teaiwa K (2024) Not Here for the Disciplines: Researching with and for the Pacific. In: Davis TC, Rae P (eds) The Cambridge Guide to Mixed Methods Research for Theatre and Performance Studies. Cambridge, University of Cambridge, pp 272–283 Tsing AL (2015) The mushroom at the end of the world: On the possibility of life in capitalist ruins. Princeton University Press, Princeton van Voorst R, Ahlin T (2024) Key points for an ethnography of AI: an approach towards crucial data. Humanit Social Sci Commun 11(1):1–5 Widder DG, Whittaker M, West SM (2024) Why ‘open’AI systems are actually closed, and why this matters. Nature 635(8040):827–833 Wilson S (2020) Research is ceremony: Indigenous research methods. Black Point. Nova Scotia, Fernwood publishing Zong J, Matias JN (2024) Data refusal from below: A framework for understanding, evaluating, and envisioning refusal as design. ACM J Responsible Comput 1(1):1–23 Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 16 Mar, 2026 Reviews received at journal 21 Jan, 2026 Reviewers agreed at journal 07 Dec, 2025 Editor invited by journal 26 Nov, 2025 Reviews received at journal 22 Sep, 2025 Reviewers agreed at journal 18 Sep, 2025 Reviewers agreed at journal 17 Sep, 2025 Reviewers invited by journal 16 Sep, 2025 Editor assigned by journal 16 Sep, 2025 Submission checks completed at journal 26 Aug, 2025 First submitted to journal 26 Aug, 2025 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-6926166","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":518680612,"identity":"561febcf-cead-487e-ad0d-ec2d004020a6","order_by":0,"name":"Bronwyn Isaacs","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABH0lEQVRIiWNgGAWjYBACAxiDDULZMDAws5GmJQ2i5QAxWqDgMEQ3Pi3m0oePPeZtu8PAx3/44Ycff84n9rezJX7+UMEgz9/Ae/gDFi2WfWnpxrxtzxjYJNKMJXvbbifOOMx2WOLAGQbDGQf4EtAdAXbYGR4z6dy2w0AtDAbSjA23cxsOszdIHGxjYNzAwGOQgFcL//HPvxn+nMudf5i9+QdQiz1ICzY/IbQw5JhJM7AdyN1wmO0YyJZEoBbDBmx+6WFLk/5z7jAPm0ROmWVvW3L9xsNsaRZnzkgkzzjMY4w1xHiYj0nOKDssJ99/fPONH3/sjOXOHzO+UVFhY9vf3mOMLcTAgJGNgQddTAIYp7jUg8AffJKjYBSMglEw4gEAGdpf5QSnE+YAAAAASUVORK5CYII=","orcid":"","institution":"University of Waikato","correspondingAuthor":true,"prefix":"","firstName":"Bronwyn","middleName":"","lastName":"Isaacs","suffix":""}],"badges":[],"createdAt":"2025-06-18 23:08:10","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6926166/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6926166/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":92185550,"identity":"cfb672b6-99f5-41c4-a123-c452118e1aff","added_by":"auto","created_at":"2025-09-25 14:10:12","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":119857,"visible":true,"origin":"","legend":"","description":"","filename":"CodeCultureArticle.MinorRevisions.26.8.25.docx","url":"https://assets-eu.researchsquare.com/files/rs-6926166/v1/aa237f5c4b39e27a87f66598.docx"},{"id":92186013,"identity":"fef3aa3e-8709-42bb-83a6-60414eb69b8f","added_by":"auto","created_at":"2025-09-25 14:18:12","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2639,"visible":true,"origin":"","legend":"","description":"","filename":"88a7548b0f0549ca898b0948bb54b0c8.json","url":"https://assets-eu.researchsquare.com/files/rs-6926166/v1/69ff741741091451c68e6ccb.json"},{"id":92185551,"identity":"2138269f-d26e-422f-b599-68b1337bbfd2","added_by":"auto","created_at":"2025-09-25 14:10:12","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":74573,"visible":true,"origin":"","legend":"","description":"","filename":"88a7548b0f0549ca898b0948bb54b0c81enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-6926166/v1/7fe1b89e687b6093ed7d5548.xml"},{"id":92187259,"identity":"cb4af820-dcaa-4f33-b9f5-d572918025ac","added_by":"auto","created_at":"2025-09-25 14:26:12","extension":"jpeg","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":316350,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6926166/v1/d29b5043e5ae7cca8b025eb4.jpeg"},{"id":92185558,"identity":"09b0c34e-c8c2-4f4c-82ff-30b8165fc6cd","added_by":"auto","created_at":"2025-09-25 14:10:12","extension":"jpeg","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":584791,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6926166/v1/abc3d5c93fa7a10d627fe81a.jpeg"},{"id":92185557,"identity":"3b8b8be9-d25b-4e13-b25b-66f33020b194","added_by":"auto","created_at":"2025-09-25 14:10:12","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":79021,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-6926166/v1/f17a42ff469ffa09848687bf.png"},{"id":92185552,"identity":"3db87955-fa1a-48f9-b26a-0696a5fedabb","added_by":"auto","created_at":"2025-09-25 14:10:12","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":221819,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-6926166/v1/7a9b1c811203e1339286da11.png"},{"id":92185556,"identity":"bf644c75-2662-4ade-bcef-6698f06d7a9a","added_by":"auto","created_at":"2025-09-25 14:10:12","extension":"xml","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":70514,"visible":true,"origin":"","legend":"","description":"","filename":"88a7548b0f0549ca898b0948bb54b0c81structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-6926166/v1/aec5cfc49065eb387e07f1e2.xml"},{"id":92185555,"identity":"d0d90b85-7a87-45cd-9eb0-dd585f6f04aa","added_by":"auto","created_at":"2025-09-25 14:10:12","extension":"html","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":78481,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-6926166/v1/0c0e85a9c8ba59fb5a750868.html"},{"id":92185548,"identity":"11f2562f-17ca-47cc-a95b-618d2d173cb4","added_by":"auto","created_at":"2025-09-25 14:10:11","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":11870,"visible":true,"origin":"","legend":"\u003cp\u003eSam Troth films content on his phone while walking on the side of the road in Huntly, Waikato, New Zealand. Photo by author.\u003c/p\u003e","description":"","filename":"1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6926166/v1/d84c9cfd6acf033476f9123c.jpg"},{"id":92185549,"identity":"e5f294e5-08c3-4c3f-99be-5efb52ab93f6","added_by":"auto","created_at":"2025-09-25 14:10:12","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":26917,"visible":true,"origin":"","legend":"\u003cp\u003eJoe Wilson prepares to interview Sam Troth in the Ngahinapouri area, near Hamilton, Waikato, New Zealand. Photo by author.\u003c/p\u003e","description":"","filename":"2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6926166/v1/02478acd1a7ab2c157f2df1c.jpg"},{"id":92187260,"identity":"c087a5fb-b179-4d1b-a81b-437040acaae3","added_by":"auto","created_at":"2025-09-25 14:26:16","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":393437,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6926166/v1/2769c748-57ea-470a-ac19-5dd93eac8569.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Using AI to Create Cultural Content: Ethnographic Insights into Everyday Ethics","fulltext":[{"header":"Introduction","content":"\u003cp\u003eThe widespread use of generative AI is quickly changing how many people think about culture and creativity. Considerable academic and public attention is directed towards various concerns, such as the distinctions between human and machine intelligence, the enhancement of productivity alongside the loss of human jobs, the manipulation of individuals through deepfakes, and the emergence of new forms of human emotional entanglement with machines. All these issues relate to culture, that is, how humans construct social worlds and webs of meaning. This paper focuses on a particular cultural issue concerning AI: what is assumed and what is lost when culture is treated as something that can be endlessly scaled up, accessed, and shared without consultation or connection with the communities with whom that culture originates.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eNew AI tools are being trained on datasets that scrape information from increasingly diverse sources. The implementation of new technologies, such as Generative Pretrained Transformers (GPT) that can both learn and fine-tune a large language model, Convolutional Neural Networks (CNNs) for image classification, Recurrent Neural Networks (RNNs) to handle complex patterns, and Generative Adversarial Networks (GANs) that produce fake images that appear real (Bengesi et al. 2024). The practice of using these technologies to work with increasingly larger data sets in evidence by all leading generative AI firms has not gone unchallenged (see Khattak, Cohen and Taylor 2024 for a list of copyright lawsuits). Moreover, the operational scales at which generative AI is created and operated also deserve deeper, culturally specific forms of ethical reflection. This study focuses on the work of small-scale video content creators in Waikato, New Zealand. The paper argues that even small-scale content creators are wrestling with deep ethical frameworks when it comes to the use of generative AI. These frameworks are embedded in localised social relationships and show how using AI to create culturally relevant content can be a deeply ethical practice.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThat generative AI research can also further ongoing colonial legacies of extraction and inequality is an ethical concern. The creation of new generative AI products has been shown to be primarily motivated by corporate financial objectives, thus shaping biases in the data collected and the kinds of racial and ethnic biases that large language models (LLMs) display (Adams 2025; Klein \u0026amp; \u0026nbsp;Ignazio 2024, Gebru 2020). Bender et al. (2021, 616) highlight that the risks of operating at the scale of generative AI\u0026rsquo;s large language models (LLM) include bias, increasing power imbalances and giving increased authority to existing hegemonic viewpoints (Bender et al 2021). Widder, Whittacker, and West (2024, 82) also highlight the power imbalances perpetuated by such models, noting that \u0026nbsp;\u0026ldquo;the present pattern in AI development takes a bigger-is-better approach when it comes to data, but. . . information on the datasets for models has become increasingly opaque, for closed and ostensibly open models alike\u0026rdquo;. This pattern of bigger is better data scraping moreover fails to do justice to the claims of data sovereignty, story sovereignty and cultural IP called for by indigenous and minority communities in different parts of the world (King, Gybele \u0026amp; Anderson 2015; Kukutai \u0026amp; Taylor 2016).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eGenerative AI changes quickly and ethnographic projects take a long time to complete. In this paper, I meet this challenge by reporting on an ongoing ethnographic project in which I am still actively engaged, with content creators and people working in the New Zealand film industry. AI has been embraced in the international film industry including in New Zealand where it is used for handling more repetitive and \u0026ldquo;boring\u0026rdquo; tasks such as background CGI work and as a tool for reflection and refining ideas (Narayan et al 2022). There are however considerable concerns and forms of resistance to AI creative technologies expressed by people working across the film production process in many parts of the world particularly regarding labour rights, copyright and IP infringement and loss of earnings (Khattak, Cohen and Taylor 2025). Using anthropologists\u0026rsquo; methodological attention to events and people in detail, in this paper I focus on a small number of content creators and their creative processes over a limited time period. I ask, how do independent media creators, working with limited resources engage with AI and think about AI in their work? Do they see their work as something which that is applicable for AI models and can be \u0026ldquo;scaled up\u0026rdquo;? And what philosophies and ethics are in play when they use or don\u0026rsquo;t use AI in their everyday creative activities?\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eEthics is understood here as the way that people as social beings wrestle with their moral obligations to others, including moments of moral breakdown and questioning (Brodwin 2013; Kleinman 2006). This presumes that ethics are not merely matters of individual reflection but can be expressed in concert with the environment and the collective in which one lives (Jackson 2011). The results of the research project discussed here indicate that people using generative AI may do so in a deeply ethical way. They may think carefully and critically about their use of AI and the risks that emerging AI technologies pose for issues of cultural respect, social relationships, and exploitative treatment of others via the use of AI technologies.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003eThe two main research methodologies on which this research project is based are ethnographic \u0026ldquo;deep hanging out\u0026rdquo; and qualitative interviews. Studies of AI ethics are often treated as \u0026ldquo;lower status\u0026rdquo; that studies in AI engineering because of the emphasis on subjective forms of knowledge (Widder 2024). Studies that consider small, scale, experiential forms of knowledge can however bring to light the views of those who are not represented in larger statistical or large scale research methodologies. This style of small scale research also meets an important call by feminist and minority researches for AI research to focuses on real people and their real experiences (Zong and Matias \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Bassett \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Klein and D\u0026rsquo;Ignazio (2024, 106) for example state, \u0026ldquo;Put simply: the field of AI needs to develop more participatory, more responsible, and more humble methods for being in dialogue with impacted communities.\u0026rdquo;\u003c/p\u003e\u003cp\u003eThis paper emerges from a project that has been in process for six months, and includes at this point, 14 interviews as well as ethnographic observation, action or participatory based research occurring over a period in of about three months. The interviewees thus far come from a range of sectors in the film industry in New Zealand, including those working in areas such as film directing, copyright, and diversity and inclusion. This study has not engaged any of the generative AI companies used or discussed by the participants. However there is research value in attending to how users use generative AI even if it is in ways not intended. As Matias (\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2023\u003c/span\u003e, 251) notes, \u0026ldquo;Companies have strong incentives to resist studies that could suggest they are responsible for harms involving their products. Therefore, future studies will depend on innovations in adversarial data access, ethics, and privacy.\u0026rdquo; The ethnographic research discussed in this paper thus focuses on small local content creators, working with small budgets and relying on self production and community support. The paper focuses on one particular day of filming which serves to centre the themes of ethics as they are experienced on the small-scale, local level. Ethical approval for this project was provided by the University of Waikato Human Research Ethics Committee, project FS2024-68.\u003c/p\u003e\u003cp\u003eEthnography\u0026rsquo;s attention the small scale, has own unique advantages to studying culture. What ethnography loses in broad applicability, Herzfeld (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2001\u003c/span\u003e:23) writes, \u0026ldquo;it gains in the sheer intensity of the ethnographic encounter- as intimacy, as privileged access, as listen to the voices silenced on the outside by those who wield great power\u0026rdquo;. Parre\u0026ntilde;as (\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) compares the work of ethnographers to moles as opposed to mining corporations. While the latter extract valuable content on a profitable and extractive scale, ethnographers who are moved by what they witness on a sensorial level are more like \u0026ldquo;tiny moles that make new connections by digging channels in the dark\u0026rdquo; (Parre\u0026ntilde;as \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e, 258). In regards to the study of AI, van Voorst and Ahlin (\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) highlight the ability of ethnography to probe things associated with our use of technologies that may be unseen or \u0026ldquo;absent-present\u0026rdquo;, things important in people\u0026rsquo;s daily experience of AI but not assumed by those who make it. Bringing these perspectives in view, this paper presumes that there is research value in focusing on a single day of filming including the experiences views of just two content creators, as is discussed in the following sections.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\n\u003ch2\u003eContent Creators Consider AI images Ethically\u003c/h2\u003e\n\u003cp\u003eAt five AM on an Autumn morning in Huntly, a semi-rural area of Waikato in New Zealand, I crunched my feet down a long gravel driveway, looking to meet three local men. These men were all working in various ways in the north island of New Zealand film, video, and content creation. These men were firstly, Joe Wilson, a local content creator and social organiser, secondly, a man working as Joe\u0026rsquo;s director of photography (DP), and thirdly, Sam Troth, a content creator and activist. Sam was currently part way through a journey of walking the length of New Zealand\u0026rsquo;s north island, a journey he saw as both self-healing from his own experiences of abuse, and a way to raise awareness of abuse issues in New Zealand. Still tucked in sleeping bags and blankets at the back of the club were Sam\u0026rsquo;s two current team members, who were also participating in the walk and assisting him with his mission. Together, the three walkers took turns walking each day, covering a route stretching from the top to the bottom of New Zealand\u0026rsquo;s North Island. Sam\u0026rsquo;s plan for this walk, which he and referred to as a \u003cem\u003ehikoi\u003c/em\u003e (a Māori word for walk or journey, often associated with a protest march) was to finish in Wellington, the country\u0026rsquo;s capital, where he and the two others walking with him would present a petition at parliament house calling for longer sentences for those convicted of sexual abuse.\u003c/p\u003e\n\u003cp\u003eJoe, a social organiser and content creator from the Waikato, was there in his capacity as the producer and director of \u003cem\u003eLots of Little Fires\u003c/em\u003e, a production company that in Joe\u0026rsquo;s words aims \u0026ldquo;to bring to light invisible labour\u0026rdquo; of people doing local charity and community development. At a previous event, Joe explained his work to me as a bridge between community workers needing support and those who can support them, particularly with funding. When I asked Joe why he makes videos, he explained, that based on prior experience of working in youth and council work, \u0026ldquo;if I write a report, nobody will read it\u0026rdquo;.\u003c/p\u003e\n\u003cp\u003eThus, despite having no background in film work, Joe had turned his advocacy approach to working with video because he viewed it as a necessary tool. Many of the videos Joe made were for local charities and community groups operating in distinct cultural contexts. This included \u003cem\u003eManaaki Rangatahi ki Waikato\u003c/em\u003e, a charity providing housing for homeless Māori youth; \u003cem\u003eRainbow Hub\u003c/em\u003e, a community group advocating for rainbow communities; and \u003cem\u003ePoutama Rites of Passage\u003c/em\u003e, an Indigenous community-led youth mentorship. Many of the videos that Joe maked through Lots of Little Fires thus engaged with content considered relevant to culture, including Indigenous culture. On the day in question, Joe, his DP and Sam met in Huntly, and Joe planned to spend the day getting footage and audio he could use to make short video that would help to highlight Sam\u0026rsquo;s advocacy and his contribution to his extended communities.\u003c/p\u003e\n\u003cp\u003eThat morning in Huntly, the night was inky black but warm under cloud cover, and we gathered outside the coast by the quiet coast motorcycle club\u0026rsquo;s concrete building. The DP took B-roll footage of Sam making a cup of coffee and checking messages on his phone while the three men (DP, Sam and Joe) joked about the amount of labour each felt compelled to do on their social media accounts. Joe explained his \u003cem\u003ekaupapa\u003c/em\u003e was to bring the stories of communities are charities to light but that hated the need to constantly post and monitor messages online. \u003cem\u003eKaupapa\u003c/em\u003e is a word from the indigenous te reo Māori language meaning purpose or vision. Stemming from Māori \u003cem\u003ete ao\u003c/em\u003e (philosophy), a kaupapa is never only for the individual. Rather the term implies social good. The term is widely used in New Zealand outside of the Maori community and allows speakers to bring a sense of ethical weight or grounding to their speech. Sam replied to Joe that his \u003cem\u003ekaupapa\u003c/em\u003e was to spread light on sexual abuse and survival. He expanded on how this was his second walk across Aotearoa to raise awareness of these issues. Sam discussed how spending hours walking process slowed life down and provided a new perspective, \u0026ldquo;almost spiritual\u0026rdquo;. This special experience also had to be filmed. Figure\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e shows a photo taken of Sam during his morning from the perspective of the van that was following him with the \u003cem\u003eLittle Fires\u003c/em\u003e film crew inside. Joe, the DP and myself spent several hours in a van, trailing beside Sam as he walked on the side of the road. Much of the time, we chose to wait before filming because we did not want footage of Sam using his phone. Sam spent quite a bit of time on his phone, calling people and creating content so that he would have something to choose to post later.\u003c/p\u003e\n\u003cp\u003eJoe\u0026rsquo;s initiative in creating promotional videos rather than writing reports, and Sam\u0026rsquo;s felt need to constantly post reflected the challenges of generating social change within an \u0026ldquo;attention economy\u0026rdquo;. By \u0026ldquo;attention economy\u0026rdquo;, scholars refer to the technological-social changes in human cognition and reliance on owing to the everyday integration of smart devices and digital communication (Paasonen, \u003cspan class=\"CitationRef\"\u003e2021\u003c/span\u003e). The attention economy as based on profit-motivated algorithms of social media platforms and the physical design of smart devices designed distract and attach users to digital interfaces, affects individuals\u0026rsquo; concentration and ability to provide sustained attention (Kozyreva et al. \u003cspan class=\"CitationRef\"\u003e2020\u003c/span\u003e). Joe operated his production company on the belief that those working for charities and local councils that give funding to community organisations in New Zealand, either do not read, or are not swayed or containing information when provided in written form, but require aesthetically appealing and emotionally resonant forms of visual media for a message to break through. Sam and Joe both discussed the need to post and use social media accounts when they would rather be doing other things.\u003c/p\u003e\n\u003cp\u003eThat morning, while drinking his pre-dawn coffee, Sam continued to discuss his need to post frequent content: It\u0026rsquo;s a lot man\u0026rdquo;, and \u0026ldquo;You\u0026rsquo;ve got to know which is the content that\u0026rsquo;s right to put up\u0026rdquo;. Going on, Sam explained, \u0026ldquo;So this week I got a call asking if I could talk to this guy who was suicidal. Talk to him, have the chat; he has been through terrible things. However, you cannot post about that. That\u0026rsquo;s not right. We could put up some pretty extreme content-stuff that would get lots of views, but we gotta do what\u0026rsquo;s right. Its like. .. \u003cem\u003eHere\u003c/em\u003e, Sam stood up walked towards the DP, his arm outstretched, \u0026ldquo;\u003cem\u003eLike - here mate here\u0026rsquo;s 20\u003c/em\u003e -- like waving 20 [dollar note] at a homeless man. That is what it would be like, taking advantage. There are all the things you cannot show.\u0026rdquo;\u003c/p\u003e\n\u003cp\u003eSam then turned his phone around, providing us with a good view of his screen which shone like a small orb against the blackness of the night around us. \u0026ldquo;There are all these things you cannot show. You must be careful when making posts depicting children in danger. I tried to make an image about a child scared in their bedroom, and AI made the child\u0026rsquo;s face a cabbage patch doll. That\u0026rsquo;s not right\u0026rdquo;. Sam continued to scroll, turning his phone between his face and ours, showing us images made by ChatGPT and other apps. Some images contained dark themes related to Sam\u0026rsquo;s work in raising awareness of abuse. There were images made in the style of black and white illustrations that depicted children hiding in their bedcovers while a shadowy figure lurked at their bedroom door. Another use of generative AI, more light-hearted in nature, was an image Sam had made in the style of a viral \u0026ldquo;doll\u0026rdquo; trend, featuring an image of a plastic-looking doll packaged as if for sale in a box. The doll resembled Sam and came with a high-vis vest, such as the one Sam wore for safety on the road, a torch and a miniature toy phone, reflective of Sam\u0026rsquo;s self-perception of his social media responsibilities while out of the road.\u003c/p\u003e\n\u003cp\u003eSam\u0026rsquo;s frustration in attempting to navigate chatGPT to produce images relevant to his \u003cem\u003ekaupapa\u003c/em\u003e is illustrative of the elisions and biases built into the LLMs on which generative AI are made. The impact of this can be considered as a form of what Kane (2014) calls the \u0026ldquo;post-optical\u0026rdquo; quality of digital interfaces and images. Using the term \u0026ldquo;post-optic\u0026rdquo;, Kane emphasises that digital images reflect the limitations of machine software more than they do the decisions of human design. Kane argued that when users create images with digital tools, they know very little about the eliminations and standardisations through which the code of images is produced. Thus, users do not determine what is seen online; rather, computer code is largely determinative of the images and thus of image culture. Kane\u0026rsquo;s argument remains extremely relevant when it comes to AI-generated images. When using generative AI to create images, the original datasets on which generative AI is created are no longer referenced directly but pertain only to a \u0026ldquo;set of distinct statistical probabilities\u0026rdquo; (Menotti \u003cspan class=\"CitationRef\"\u003e2025\u003c/span\u003e, 2). The images Sam produced of scared children therefore pertained to no individual instance in which an artist considered themes of children and fear, but a statistical formula, the parameters of which were standardised and invisible.\u003c/p\u003e\n\u003cp\u003eThe day of filming continued and was full of creative improvisation and hustle. The theme of the day\u0026rsquo;s filming from my perspective as an outsider seemed to be getting things done with minimal resources, and making content with what was available. For example, Joe had a few ideas of locations that day where we could film. However, on arriving at one of the parks Joe had planned on, we were unsuccessful in finding a quiet spot to film. Our filming date was during school holidays, and the park was busy with foot traffic. Joe, the DP and I discussed among ourselves possibly alternatives such as filming in various people\u0026rsquo;s back gardens and other parks in town. Eventually myself, Joe and the DP drove around the rural area of Ngahinapouri, in order to find a location quiet and with a visually striking backdrop in order to film the interview with Sam, finally settling on the location in Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e.\u003c/p\u003e\n\u003cp\u003eFor Sam, in the terms he described and used, the use of generative AI was a tool in his everyday social media practices of posting content to social media. These posts were made, as Sam explained, to raise awareness and connect with others on a similar journey of healing. While this work was important to him, Sam felt exhausted by it, demonstrating a \u0026ldquo;posting fatigue\u0026rdquo; that is not uncommon in attention economies (Bright, Klesier \u0026amp; Grau 2015). Using generative AI to create images for posts was not necessarily core to Sam\u0026rsquo;s \u003cem\u003ekaupapa\u003c/em\u003e, but as it was a way, when his resources were limited, of creating content, a task he felt to be urgent and necessary.\u003c/p\u003e\n\u003cp\u003eLater that day, Joe interviewed Sam about why he spent so much time and energy raising awareness about sexual abuse. Sam expanded on his \u003cem\u003ekaupapa\u003c/em\u003e, explaining he was doing healing work on himself and doing something good for the world and his kids. Joe turned to me and to his DP several times throughout the interview asking, \u0026ldquo;Is this getting the story arc? Is there a question I should be asking?\u0026rdquo;. Here, a human, face-to-face discussion was valuable owing the gravity of what was being discussed and the open-endedness of the creative process. Throughout the day of filming, we discussed our connection to and the necessary labour for creating content on social media. In this moment, however, the kaupapa of both Joe and Sam as content creators came through in their deeply social commitments.\u003c/p\u003e\n\u003c/div\u003e\n\u003ch3\u003eContent Creators Care about Creativity\u003c/h3\u003e\n\u003cp\u003eWithin a few weeks I followed up with Joe to discuss his editing process for the footage and audio that Joe and the DP collected on the day of the shoot with Sam. Perhaps surprisingly, Joe explained to me that the goal with videos was not maximum reach or creating viral videos, but instead his goal in posting the videos was \u0026ldquo;not how many people saw the video but if the right person did\u0026rdquo;. Explaining the importance of making connections with people, Joe said emphasised filming days were typically short (usually, as with the film about Sam, only one day of filming), but that they were done in the context of ongoing relationships: \u0026ldquo;For me, it\u0026rsquo;s so important to do those earlier visits so that people get used to you, but it can be really challenging for many people to be comfortable on camera. I have all these questions written down I mean to ask them but I end up chatting about all kinds of other things.\u0026rdquo;\u003c/p\u003e\n\u003cp\u003eJoe viewed the power of what he did in communicating with others through the lens of storytelling. \u0026ldquo;I come from the North of England you see\u0026rdquo;, Joe explained to me,\u003c/p\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eWhere I come from, we have a tradition of folk stories, of telling stories that speak to people. Here, Māori stories also have power. And as a songwriter, I see the power of story vulnerability. When you are busking, you see people respond to those songs where you share a piece of yourself.\u003c/p\u003e\n\u003c/div\u003e\n\u003cp\u003eThese stories explained Joe, are communicated most powerfully in small-scale settings.\u003c/p\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eThat\u0026rsquo;s why I call my program \u003cem\u003eLots of Little Fires\u003c/em\u003e, because if you scale up you lose the campfire - you lose social connections, the fire needs to be done on a smaller scale. A big fire is too hot, and a big social project likewise results in lost relationships and innovation. I worked with the youth for the local council. I was very frustrated by the lack of care they had for people. But if you get people one or one, or in a small group, then they see each other as people, they can hear them.\u003c/p\u003e\n\u003c/div\u003e\n\u003cp\u003eBased on this philosophy of social connection, Joe carefully manages who has access to his filming process. In choosing a videographer to work with, Joe also emphasised that he did not look for a particular skill set or aesthetic style, but rather that the videographer needed to \u0026ldquo;be humble\u0026rdquo; because \u0026ldquo;the quality is only as good as the work and the people\u0026rdquo;. In regard to my presence at events, Joe judged, for example, that my presence at a meal-sharing service that he planned to film may lead some people to feel uncomfortable. At a free meal service, the costs of participants feeling fear of judgement and possible recognition from outsiders were high. However in regards to the filming activity discussed in this paper, that is, the filming of Sam Troth, who was himself content creator and keen to get more attention for his channel, Joe checked and learned that Sam would happy for me to follow the filming process.\u003c/p\u003e\n\u003cp\u003eThis focus on personal and social relationships as key to negotiating the ethics of content creation was also evident in Joe\u0026rsquo;s selective use of AI tools in his work. On the one hand, Joe quickly accepted his videographer\u0026rsquo;s recommendation of using AI tool REV to take audio files of filming content and transcribe them. When I asked Joe about this he said that the DP sent him the interviews as a REV file and this \u0026ldquo;saves a huge amount of time, work that we would have to do\u0026rdquo;. For Joe, however, the use of generative AI demands careful, ethical reflection. Speaking of his videographer and his advice regarding AI, Joe commented, \u0026ldquo;[The DP] has said that we could put the interviews into ChatGPT and get ChatGPT to select the interview content along with the timestamp and that could be really useful but I do not want to go down that road. Because the work we do is really relational and I do not want to give to chat GPT and let it die.\u0026rdquo;\u003c/p\u003e\n\u003cp\u003eWhen I asked what he meant by \u0026ldquo;let it die\u0026rdquo;, Joe clarified, his concern was twofold,\u003c/p\u003e\n\u003cp\u003e\u0026ldquo;It\u0026rsquo;s ok to use AI to save time but I don't want artificial intelligence doing the creative work that humans should be doing. Artificial intelligence should not be doing the creative work of humans. And what the people I work with give me is a gift. It\u0026rsquo;s their stories, their suffering and my job to honour that. Its not that we shouldn't use AI, but we must use it in the right way. For me, it\u0026rsquo;s the story teller tradition - connecting with people first and foremost\u0026rdquo;.\u003c/p\u003e\n\u003cp\u003eJoe\u0026rsquo;s concerns with AI thus fell into two categories, protecting the category of human creativity and secondly, honouring the sharing of stories as a gift \u0026ndash; something outside of simple transactional exchange. Joe\u0026rsquo;s concerns about human creativity as distinct from machine-generated content are reflected in wider sectors of content and film industry (Lee \u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e). Khattak, Cohen, and Taylor (\u003cspan class=\"CitationRef\"\u003e2025\u003c/span\u003e, 2) note that the issue of creativity as a human-led process is particularly fraught in the film industry, where \u0026ldquo;artists, screenwriters, and filmmakers have increasingly pushed back, warning that AI threatens not only their jobs but also the core of artistic authorship and cultural integrity\u0026rdquo;. Joe\u0026rsquo;s unwillingness to give data to ChatGPT is also an example of what some scholars have identified as the practice of data refusal. Zong and Mattias (2024) for example, who draw on the work of Ruha Benjamin\u0026rsquo;s (\u003cspan class=\"CitationRef\"\u003e2019\u003c/span\u003e) work on the relationship between technology and racism, defining data refusal as a practice on practice used by those whose lives may seem peripheral to the centre of tech decision making but who are directly impacted by its design. Refusing to engage in acts of data collection can be for those on the periphery of tech worlds, an act of \u0026ldquo;contest[ing] the power of data in their lives, while also making the power of data collectors more legible\u0026rdquo; (Zong and Mattias 2024 3).\u003c/p\u003e\n\u003cp\u003eJoe\u0026rsquo;s use of the term \u0026ldquo;gift\u0026rdquo; moreover reflects a deeply moral position on what is distinctly human. Gifts are a key practice in human culture. Anthropologists have shown gift giving to be possibly universal and certainly an ancient and enduring form of social obligation, binding the giver to the receiver in a debt of obligation (Graeber, \u003cspan class=\"CitationRef\"\u003e2011\u003c/span\u003e). Gifts can also frequently have a sacred quality, as they can entail giving a part of oneself to another (Mauss 1925). By linking the gift of a story to a wider tradition of storytelling, Joe\u0026rsquo;s ethical perspective also reflects a broader intentional ethical and epistemological view of knowledge sharing and storytelling. In this epistemology, as practiced in many Indigenous and other minority communities, as eloquently outlined by scholars such as Frederick (2003), Wilson (\u003cspan class=\"CitationRef\"\u003e2020\u003c/span\u003e), Teaiwa (\u003cspan class=\"CitationRef\"\u003e2024\u003c/span\u003e), and Funaki (2025), knowledge shared in the context of community and social relationships holds a strong ethical quality that binds social ties and can increase capacity for shared responsibility and collective action.\u003c/p\u003e\n\u003cp\u003eCurious as to whether Joe\u0026rsquo;s concerns regarding AI, the safeguarding of the \u0026ldquo;gift\u0026rdquo; of shared stories reflected more broadly with the small communities and charities he works with, I asked Joe, \u0026ldquo;Do the charities and communities share your perspective on AI or do not really care?\u0026rdquo;. Joe replied, \u0026ldquo;Its not that they don't care. It\u0026rsquo;s that they may choose to use generative AI when they need an output and don't have many resources. They don't have the support and resources they need, so they may choose to rely on it\u0026rdquo;.\u003c/p\u003e\n\u003cp\u003eContesting or refusing data extraction is therefore an example of Joe expressing his \u003cem\u003ekaupapa\u003c/em\u003e, his ethical commitment, and purpose. As Joe eloquently explained, the metaphor of a little fire is that a fire that is too large is too hot. Here it is not size per see that is the issue, but rather the dangers of lost social connections and respectful engagement of others around the metaphorical fire because there is value in paying attention to how social connections alter at different scales. To be clear, the issue is not with scale per see, but in expansion without reference to its social and cultural impact. This point is forcefully made by Tsing (\u003cspan class=\"CitationRef\"\u003e2015\u003c/span\u003e), who argues that many capitalist projects from plantation slavery to computing attempt to reduce complex and diverse information about people and the environment to comparable units. A computing-based scalable unit such as a pixel is designed to be endlessly replicable, allowing for a potentially endless process of zooming, vision going in, out, and across. Such scalability explains that Tsing (\u003cspan class=\"CitationRef\"\u003e2015\u003c/span\u003e, 38) \u0026ldquo;banishes meaningful diversity\u0026rdquo;. Computing based attempts to treat things as uniform without paying attention to local questions of culture or ethics risk replicating colonial models of expansion and extraction and inform the most extractive versions of contemporary or late capitalism.\u003c/p\u003e\n\u003cp\u003eIn the context of the work of small content creators Joe and Sam, generative AI was difficult to ignore owing to the opportunity it offered to decrease labour time. As content creators, Sam and Joe were not operating according to traditional capitalist models in that they were not seeking commercial funding or payments for their products. Nonetheless, the broader patterning of the attention economy has impacted the media industry in which their work was created. Even within the attention economy, however, the creative process, including that of AI use, was navigated by Sam and Joe through a committed ethical framework. For Joe, it was about the integrity of the small community, where people know and understand each other. For Sam, a larger network was sought, but it had to be handled on a person-to-person basis. For both, it was important to allow for individual suffering and circumstances to shape the ethics of how content is created and treated.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIn attending to the small scale, this paper focused on ethical and epistemological arguments for better understanding both the limits and harms of AI in regard to representation, study, and creative engagement with culture. Content creators such as Sam and Joe view AI as a tool to save time on mundane tasks but are cautious about using it for creative work or handling sensitive stories. This research project should not be assumed to have broad statistical replicability. Nonetheless, in paying attention to what is important to local content creators, this research demonstrates the importance of considering ethical concerns as part of users\u0026rsquo; experiences. This research project provides evidence of content creators balancing the efficiency of generative AI with ethical concerns pertaining to preserving human creativity and respecting cultural content. The concept of \"kaupapa\" (purpose/vision) is important to these content creators' ethical decisions regarding AI use.\u003c/p\u003e\u003cp\u003eAccording to the \u0026ldquo;bigger is better\u0026rdquo; models of AI design, there are few limits to the kinds of media and content which future datasets might access. There are however possible alternatives to this presumed model of continued extractive data scraping. Legally, it is appropriate to consider increased protection for individual privacy, copyright, and IP holders, in addition to cultural sovereignty over community cultural content. For tech companies engaging in generative AI data training, there is also the possibility of \u0026ldquo;algorithm abandonment\u0026rdquo; recognising the social harms of existing algorithms and when needed, agreeing to dispense with them (Johnson et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e\u003cp\u003ePrimarily, this paper calls for researchers in academia and technology to better appreciate the way in which AI users use AI with moral and ethical intentionality. AI users are not replicable standardised units. They are complex and diverse owing to their culture, their ethics and commitments to their communities.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData Availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data was obtained through informal interviews and participant observation. It cannot be shared in accordance with the ethical procedures in which this study was conducted, including respect for participants and the sensitivity of their personal experiences. While the main participants mentioned in the ethnography chose not to be anonymous, they did not give permission for all interview data and research notes to be shared.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eEthical approval for this project was provided by the University of Waikato Human Research Ethics Committee, project FS2024-68 on 16 December 2024. The procedures used in this study adhered to the principles of the Declaration of Helsinki.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003eInformed Consent:\u003c/h2\u003e\n\u003cp\u003eInformed consent to participate in the study was obtained from all participants on 18 March 2025 and 15 April 2025 respectively. Participants received an information sheet outlining the nature and aims of the research and were given the opportunity to ask questions prior to providing consent. Consent was documented via signed consent forms; in cases where participants preferred, verbal consent was obtained and recorded by the researcher. All participants were offered the option to remain anonymous. The individuals discussed directly in this article are content creators engaged in advocacy work and have chosen to be identified.\u003c/p\u003e\n\u003ch2\u003eFunding:\u003c/h2\u003e\n\u003cp\u003eFunding was provided by the University of Waikato. Funding was provided directly to the author\u0026rsquo;s individual research trust account. This funding was provided for research purposes. Beyond this there is no particular name or grant associated with the funding.\u003c/p\u003e\n\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\n\u003cp\u003eB.I. did the research and wrote the manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAdams R, Adeleke F, Junck L, Alayande A, Segun S, Gupta A, Zaman M et al (2025) Mapping the Potentials and Limitations of Using Generative AI Technologies to Address Socio-Economic Challenges in LMICs. VeriXiv 2:57. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.12688/verixiv.948.1\u003c/span\u003e\u003cspan address=\"10.12688/verixiv.948.1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBassett C (2023) The Cruel Optimism of Technological Dreams. In: Jude Browne (ed.), Stephen Cave (ed.), Eleanor Drage (ed.), Kerry McInerney (ed.) Feminist AI: Critical Perspectives on Algorithms, Data, and Intelligent Machines, p 258\u0026ndash;273\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBender EM, Gebru T, McMillan-Major A, Shmitchell S On the dangers of stochastic parrots: Can language models be too big?列. In: Proceedings of the 2021 ACM conference on fairness, accountability, and transparency, pp. 610\u0026ndash;623\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBengesi S, El-Sayed H, Sarker MK, Houkpati Y, Irungu J, Oladunni T (2024) Advancements in Generative AI: A Comprehensive Review of GANs, GPT, Autoencoders, Diffusion Model, and Transformers. IEEe Access. 69813\u0026ndash;69837\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBenjamin R (2019) Captivating technology: Race, carceral technoscience, and liberatory imagination in everyday life. Duke University Press, Durham\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBright LF, Kleiser SB, Grau SL (2015) Too much Facebook? An exploratory examination of social media fatigue. Comput Hum Behav 44:148\u0026ndash;155\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBrodwin P (2013) Everyday ethics: Voices from the front line of community psychiatry. Univ of California, Berkeley\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFrederick Marla (2003) Between Sundays: Black women and everyday struggles of faith. Univ of California, Berkeley\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFunaki H, Macdonald L, Knox J, McKinnon D (2025) Living in the Telling: Indigenous storytelling of post covid desires for academia. Art Res Int Transdisciplinary J 8(2):499\u0026ndash;518\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGebru T (2020) Race and gender. The Oxford handbook of ethics of AI, 4, 253\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGraeber D (2011) Debt: The First 5,000 Years. Melville House, London\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHerzfeld M (2001) Anthropology: Theoretical practice in culture and society. Malden, Blackwell Publishers\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJackson M (2011) Life within limits: Well-being in a world of want. Duke University Press, Durham\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJohnson N, Moharana S, Harrington C, Andalibi N, Heidari H, Eslami M (2024) The Fall of an algorithm: Characterizing the dynamics towards abandonment. In Proceedings of the 2024 ACM Conference on Fairness, Accountability, and Transparency, p 337\u0026ndash;358\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKhattak M, Cohen SE, Taylor K (2025) Who Holds the Camera? Filmmaking Justice in the Era of Generative AI. GRACE: Global Review of AI Community Ethics, 3(1)\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKing L, Gubele R, Anderson JR (2015) Survivance, Sovereignty, and Story. Denver, University Press of Colorado\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKlein L, D'Ignazio C (2024) Data feminism for AI. In Proceedings of the 2024 ACM Conference on Fairness, Accountability, and Transparency, June 2024, p100-112\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKleinman A (2006) What really matters: Living a moral life amidst uncertainty and danger. Oxford University Press, Oxford\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKozyreva A, Lewandowsky S, Hertwig R (2020) Citizens versus the internet: Confronting digital challenges with cognitive tools. Psychol Sci Public Interest 21(3):103\u0026ndash;156\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKukutai T, Taylor J (2016) Indigenous data sovereignty: Towards an agenda. ANU, Canberra\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLee HK (2022) Rethinking creativity: creative industries, AI and everyday creativity. Media Cult Soc 44(3):601\u0026ndash;612\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMatias JN (2023) Humans and algorithms work together\u0026mdash;so study them together. Nature 617(7960):248\u0026ndash;251\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMauss M (2024) [1925]). The Gift: the form and reason for exchange in archaic societies Oxon. Routledge\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMenotti G (2025) The model is the museum: Generative AI and the expropriation of cultural heritage. AI \u0026amp; SOCIETY, pp 1\u0026ndash;5\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNarayan AD, Caillard D, Matthews J, Nairn A (2022) Artificial imagination: Industry attitudes on the impact of AI on the visual effects process. Interactions: Stud Communication Cult 13(2):113\u0026ndash;131\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePaasonen S (2021) Dependent, distracted, bored: Affective formations in networked media. MIT Press, Cambridge\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eParre\u0026ntilde;as JS (2023) Ethnography after anthropology: Become moles, not mining corporations. Am Ethnologist 50(3):453\u0026ndash;461\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTeaiwa K (2024) Not Here for the Disciplines: Researching with and for the Pacific. In: Davis TC, Rae P (eds) The Cambridge Guide to Mixed Methods Research for Theatre and Performance Studies. Cambridge, University of Cambridge, pp 272\u0026ndash;283\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTsing AL (2015) The mushroom at the end of the world: On the possibility of life in capitalist ruins. Princeton University Press, Princeton\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003evan Voorst R, Ahlin T (2024) Key points for an ethnography of AI: an approach towards crucial data. Humanit Social Sci Commun 11(1):1\u0026ndash;5\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWidder DG, Whittaker M, West SM (2024) Why \u0026lsquo;open\u0026rsquo;AI systems are actually closed, and why this matters. Nature 635(8040):827\u0026ndash;833\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWilson S (2020) Research is ceremony: Indigenous research methods. Black Point. Nova Scotia, Fernwood publishing\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZong J, Matias JN (2024) Data refusal from below: A framework for understanding, evaluating, and envisioning refusal as design. ACM J Responsible Comput 1(1):1\u0026ndash;23\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"humanities-and-social-sciences-communications","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"palcomms","sideBox":"Learn more about [Humanities \u0026 Social Sciences Communications](http://www.nature.com/palcomms/)","snPcode":"41599","submissionUrl":"https://submission.springernature.com/new-submission/41599/3","title":"Humanities and Social Sciences Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-6926166/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6926166/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis paper explores how small-scale content creators in New Zealand navigate the ethics of using artificial intelligence (AI) in their everyday work. Ethnographic methods including \"deep hanging out\" and qualitative interviews, were used to study content creators. The paper makes a number of findings. First, users of AI technology may approach it with moral and ethical intentionality, rather than as passive consumers. Second, there is tension between the pressure to create content for the \"attention economy\" and to maintain ethical practices, especially when dealing with culturally sensitive material. Third, small-scale, relationship-based content creation is valued for maintaining social connections and respecting cultural contexts. This paper highlights the importance of considering scale and everyday user experience when considering the ethics of generative AI.\u003c/p\u003e","manuscriptTitle":"Using AI to Create Cultural Content: Ethnographic Insights into Everyday Ethics","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-25 14:09:47","doi":"10.21203/rs.3.rs-6926166/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-03-16T06:49:01+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-01-21T06:25:55+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"16425584305597202895324357833744049866","date":"2025-12-07T21:31:05+00:00","index":"hide","fulltext":""},{"type":"editorInvited","content":"","date":"2025-11-26T18:34:24+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-09-22T11:08:31+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"291908709121667187616740023516950129917","date":"2025-09-18T05:02:54+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"318349333334737171692474000468353400506","date":"2025-09-17T08:52:17+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-09-17T01:45:52+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-09-16T10:41:09+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-08-26T10:10:27+00:00","index":"","fulltext":""},{"type":"submitted","content":"Humanities and Social Sciences Communications","date":"2025-08-26T10:07:52+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"humanities-and-social-sciences-communications","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"palcomms","sideBox":"Learn more about [Humanities \u0026 Social Sciences Communications](http://www.nature.com/palcomms/)","snPcode":"41599","submissionUrl":"https://submission.springernature.com/new-submission/41599/3","title":"Humanities and Social Sciences Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"a09e7a86-cc47-4f71-92eb-d40dc41af6db","owner":[],"postedDate":"September 25th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":55114087,"name":"Social science/Anthropology/Social anthropology"},{"id":55114088,"name":"Social science/Science technology and society"}],"tags":[],"updatedAt":"2026-04-29T20:23:13+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-25 14:09:47","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6926166","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6926166","identity":"rs-6926166","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}