Exploring Indigenous personas with GenAI: A pilot study
Generative artificial intelligence is increasingly being embedded in higher education, yet its use in relation to Aboriginal and Torres Strait Islander knowledges raises significant ethical questions. This pilot study examined a Blackboard Ultra AI Conversation role-play activity autogenerated from the scenario “I would like to design an Indigenous Knowledge education module”. The activity generated an AI persona, “Aunty Nira Thompson”, positioned as a cultural advisor. A transcript of ten interactions was analysed using the AIATSIS Guide to evaluating and selecting education resources, alongside an ethical framework for generative AI and Indigenous higher education. Findings showed that the chatbot often used respectful language, recognised limits around knowledge sharing, and discouraged inappropriate use of AI-generated cultural stories and images. However, these safeguards were insufficient. The bot simulated Indigenous cultural authority without Indigenous authorship, endorsement, attribution, co-design, or governance. It also generalised Indigenous knowledges and displaced responsibility for cultural legitimacy onto future Indigenous consultation. The study argues that ethical GenAI use in this area cannot rely on disclaimers or better prompting alone. Indigenous-led governance, attribution, consent, and co-design are necessary if educational technologies are to support rather than undermine cultural authority.
Authors
- Kylie Day (ORCID: https://orcid.org/0000-0003-0202-7808)
- Zachery Quince (ORCID: https://orcid.org/0000-0003-0650-2821)
- Aimee Andersen (ORCID: https://orcid.org/0009-0001-6929-1120)
Institutions
- Southern Cross University (AU)
Publication Details
- Journal
- ASCILITE Publications
- Published
- 2026-11-17
- DOI
- https://doi.org/10.65106/apubs.2026.3226
- Primary Topic
- Artificial Intelligence in Education
- Type
- article
- Field-Weighted Citation Impact
- 0.00