The advent of AI technologies, such as OpenAI’s ChatGPT and DeepSeek, has impacted various research fields, including social sciences . AI is now widely used for generating text, images, videos, and data analysis, and its algorithms analyze user behavior to provide personalized content . In visual anthropology and sociology, AI functions as a content generator, detector, and humanizer, transforming the language and landscape of research (Khattri et al., 2025).
Historically, visual research methods were rooted in the indexical relationship of trace to reality, assuming visual objects remained largely intact . However, AI challenges this fundamental premise by offering new capabilities, such as extracting patterns from faces and bodies that humans cannot perceive . While AI can enhance human visual analysis, especially with large volumes of data, it also introduces significant ethical questions regarding the authenticity and cultural context of interpretations .
Digital ethnography, a powerful extension of visual anthropology and sociology, now integrates AI to combine visual, textual, and interactive data from various digital sources . Examples include AI systems achieving high accuracy in archaeological research by flagging potential sites for human verification, significantly reducing human labor . This collaborative approach, termed ‘augmented intelligence,’ enhances human capabilities rather than replacing them .
However, the integration of AI into visual research presents a ‘double burden’ from an ethical perspective, adding to the already complex ethical terrain of misinterpretation and commodification of cultural moments . This raises key questions such as whether researchers are satisfied with AI’s interpretation authenticity, if AI can truly grasp the depth of visual details, and what ethical issues arise from AI’s pervasive support .
Authenticity and interpretation
A major concern is the authenticity of AI-generated interpretations. AI-generated content may not reflect the cultural background of researchers or audiences, actively reshaping how researchers categorize and make sense of visual data . Over-dependency on AI can lead to false conclusions, as demonstrated by algorithms failing to recognize culturally specific visual markers or exhibiting biases against certain demographics . The use of AI-generated imagery, like DeepFakes, also makes it difficult to distinguish original content from manipulated visuals, raising serious ethical dilemmas regarding knowledge status .
Intellectual property and data privacy
The widespread integration of AI features into software makes it difficult for individuals to avoid AI use . Transparency is crucial, requiring clear guidelines on how AI has been used in written work . A significant risk is that AI programs can use submitted manuscripts to train their models, potentially putting the content into the public domain . Using AI as a reviewer also breaches intellectual property rights and may violate data protection acts if manuscripts contain personally identifiable data .
Informed consent and data sovereignty:
Integrating AI into social research introduces new challenges regarding informed consent and data sovereignty . ‘Data sovereignty’ emphasizes communities’ rights to control the collection and analysis of their visual representations, which can be compromised by AI use, especially concerning indigenous communities with specific cultural protocols .
Moral responsibility and bias
AI systems can create a ‘moral crumple,’ where responsibility for actions becomes diffuse, making it challenging to maintain ethical accountability . Algorithmic work identifies patterns that influence research conclusions, raising questions about who is responsible for interpretations . Furthermore, AI biases can systematically distort analysis, particularly when dealing with marginalized communities, potentially leading to ‘algorithms of oppression’ that perpetuate biases related to gender, color, ethnicity, race, and disability .
The need for ethical frameworks
Given the increasing necessity of AI technologies in visual anthropological and sociological research, establishing ethical frameworks is paramount . The paper emphasizes the need for AI-specific ethical codes to safeguard participants, preserve cultural sensitivity, and ensure research integrity .
Key principles for ethical AI use, summarized by Rose (2016) based on ESRC guidelines, include voluntary participation, doing no harm, providing appropriate information, maintaining anonymity, review, and avoiding conflicts of interest . However, applying these principles to visual research is complex because individuals and places are often identifiable in visual data . This raises questions about the specific ethical rules for visual anthropological sociological research, how to measure ethical standards, and how to maintain anonymity while using AI for editing ethnographic film .
The paper advocates for co-producing ethical codes with participants and practitioners to uphold the integrity of visual ethnography in the age of AI . It also stresses that technical literacy is becoming a prerequisite for ethical practice beyond computer science . The complementary roles of human and algorithmic analysis in research findings are undeniable, making ethical practice in AI, especially in visual anthropology and sociological research, more important than ever .
