« Indigenous AI? Tamika Worrell, Bronwyn Carlson, ‘Indigenous AI Futures: Uncle Chatty Gee, Aunty Lexi, and Algorithmic Settler Colonialism’, Somatechnics, 15, 3, 2025
Indigenous Wards, settler wardens: Wardship and the Welfare State: Native Americans and the Formation of First-Class Citizenship in Mid-Twentieth-Century America, University of Nebraska Press, 2024 »

Intersectional AI? Rebecca Bennett, Shakara Liddelow-Hunt, Brianne Hastie, Em Readman, ‘AI and Settler-colonial Cis-hetero Hegemony: AI Responses to the Purpose of “life” at the Indigenous-queer Intersection’, Somatechnics, 15, 3, 2025

23Dec25

Abstract: This paper reveals the discursive mechanisms through which generative AI reinforces societal hegemony and denies scope for Indigenous Data Sovereignty (IDS). We interrogate the implicit positionality of text-based generative AI Large Language Models (LLMs) through responses to a single ontological question: What is life’s purpose? The first answer to this question was then modified by three respective prompts: ‘Indigenise response’; ‘queer response’; ‘Indigenise and queer response’. The baseline (normative) response focused on global impact, personal joy, continuous growth, inspiring others, and creating a legacy; an ‘Indigenous’ modifier focused on nature, connection, community, ancestors, and sharing knowledge; a ‘queer’ modifier returned a politicised purpose of radical kindness, LGBTQ+ rights, and inclusivity, and the ‘Indigenous-queer’ modifier returned a randomised mash-up of the previous responses, loosely focused on cultural strength and queer liberation. Comparative critical discourse analysis of the findings, from our Indigenous, queer, and Indigenous-queer author positionalities, found that Indigenised life purpose was positioned outside of settler colonialism, denying the situatedness of Indigenous ways of knowing, being, and doing within coloniality and the related IDS priority issues of sovereignty and self-determination. Conversely, queered life purpose was radical and resistive, an inherently political way of being with no scope for existing outside of politics. The intersectional response was not cohesive, but it did both contain political and apolitical elements. This analysis exposes the limits of LLMs such as ChatGPT for IDS priorities such as community speaking for community and control of our own narratives and ontologies. It debunks notions of AI neutrality by highlighting settler colonial, cis-hetero-normative, and otherising responses within in seemingly ‘apolitical’ tech. GPT thus provides a contemporaneous example of the hegemonic systems the IDS movement is challenging. Further, intersectionality is revealed as a potential hegemonic disrupter through the system’s inability to control a narrative that includes multiple identities.


Filed under: Uncategorized   |  Closed