A/HRC/60/66 92. Artificial intelligence technologies, such as machine learning and natural language processing, can be used to preserve Indigenous languages, oral histories and cultural practices. Projects that digitize and document Indigenous knowledge can help to safeguard it for future generations but must be implemented with Indigenous consent and governance to prevent misuse or exploitation. 93. With the increasing growth of data connectedness, Indigenous data protection, and data sovereignty maintenance for Indigenous Peoples, become key challenges. Although artificial intelligence technologies are relatively new, they remain reliant on already-existing data – data whose narratives have historically excluded, erased, stereotyped and invalidated Indigenous Peoples, their knowledge and their voices.106 94. The digitization of Indigenous data must be compliant with Indigenous Peoples’ right to self-determination and data governance; free, prior and informed consent; their right to participate in decision-making; and their privacy and intellectual property rights. Implementation of due diligence principles, good faith consultations and free, prior and informed consent would create space and opportunities for Indigenous Peoples to organize and govern data. Indigenous Peoples exercise self-determination to identify and eliminate gaps by fostering community-driven data collection.107 95. Artificial intelligence technologies must respect Indigenous data sovereignty and incorporate Indigenous data governance principles to avoid perpetuating colonial narratives. Artificial intelligence can either perpetuate colonial narratives or serve as a revolutionary tool for Indigenous Peoples. If artificial intelligence is developed and used without accounting for the lived realities of Indigenous Peoples and their world views, the technology will reflect, perpetuate and potentially exacerbate existing inequities.108 96. For instance, the information technology developer community is reflected in the algorithmic biases of artificial intelligence-based products and services. As they are made by a minority, members of which have similar backgrounds and views of the world, and use data sets biased by those views, their products are not representative.109 97. In New Zealand, Māori are underrepresented in government data, which means that the use of artificial intelligence would impact Māori disproportionately. Supermarkets in the country have introduced facial recognition systems, and there have already been cases of Māori being misidentified.110 98. The real social power wielded by algorithms lies in their ability to determine and reorganize the boundaries of practices such as property ownership. For instance, the land records digitization programme of India dispossessed Indigenous Adivasi communities by failing to account for traditional practices of collective property rights, which defy dominant notions of private and individual ownership. This pattern is also visible in Nagaland, where carbon datafication obscured communal understanding of lands, allowing for land management practices that serve carbon finance goals, resulting in the material reorganization of space and in erosion of Indigenous access to and control over Naga land and resources.111 99. With regard to emerging digital and technology issues, there are private projects involving emerging technologies for conservation activities and sustainable investments. These projects collect, digitalize and archive information concerning Indigenous Peoples and their territories (satellite images, and audio and video documentaries) without any protection for the rights of Indigenous Peoples over their intellectual property or data sovereignty.112 106 107 108 109 110 111 112 GE.25-12012 Ian Falefuafua Tapu and Terina Kamailelauli’i Fa’agau, “A New Age Indigenous instrument: artificial intelligence and its potential for (de)colonialized data”, Harvard Civil Rights-Civil Liberties Law Review, vol. 57 (2022). See https://www.youtube.com/watch?v=2Z_SFE_rSyI. Tapu and Fa’agau, “A New Age Indigenous instrument”. UNESCO, Inteligencia artificial centrada en los Pueblos Indígenas: perspectivas desde América Latina y el Caribe (2023) (in Spanish). Presentation by Tahu Kukutai, expert seminar, December 2024. Submission from IT for Change. A/HRC/54/31, para. 60. 15