Emma Garlett: Artificial intelligence presents a profound and complex challenge to Indigenous communities

The rapid, almost vertical global adoption of Artificial Intelligence — a surge reflected in Australia’s own 45.6 per cent usage rate — presents a profound and complex challenge to Indigenous communities. The research highlights a vital duality: AI is a powerful tool for cultural resurgence, yet it also poses risks to simultaneously be a vehicle for a new form of digital appropriation. The key to navigating this future, and truly realising the productivity uplift is not just in a national strategy, but in fundamentally embedding Indigenous Data Sovereignty and the CARE Principles into the very architecture of AI development and into new necessary laws. On one hand, we all know the benefits of AI are tangible and transformative. Improving productivity, transformation and automation. It can also offer benefits in applications such as cultural protection and environmental management. The digitisation of endangered languages, transcription of oral traditions are acts of cultural and economic self-determination. It represents an opportunity to leapfrog decades of historical disadvantage, using modern tools to safeguard ancient knowledge for younger generations. However, the risks of appropriation are far more insidious and, if unchecked, could eclipse the benefits. The core issue is that AI models are trained on vast, often unvetted and uncontextualised datasets scraped from the internet. When this corpus includes Indigenous cultural, intellectual, and artistic property, the resulting AI-generated output — be it images, stories, or “styles” of art — becomes an act of theft without consent, attribution, or compensation. There are also risks. The application of AI in high-stakes social systems, such as predictive policing or child protection, poses a grave threat. If AI models are trained on historically biased data this will risk amplifying systemic racism, mistaking surveillance for safety, and causing disproportionate harm to already marginalised people and communities. The recently published Australian National AI Plan must move beyond simply capturing economic opportunities — it must prioritise the “keeping Australians safe” mandate by addressing these deep-seated equity issues head-on, especially related to First Nations peoples. Ultimately, the intersection of AI and Indigenous communities cannot be managed through existing, general laws, because the colonial history has ensured an inherent power imbalance. The answer lies in Indigenous Data Sovereignty. The CARE Principles: Collective Benefit, Authority to Control, Responsibility, and Ethics — are the essential guardrails. They demand that Indigenous peoples are not just consulted, but have the authority to control how their data is collected, stored, and used, ensuring all AI applications result in collective benefit. For AI to be a tool of empowerment and resurgence, rather than a new front of colonisation, Indigenous communities must lead its design and governance. The challenge for Australia’s government, corporations, and the 49 per cent of the population using AI is to recognise that the ethical development of this technology is inseparable from upholding the rights of its First Peoples. Emma Garlett is a legal academic and Nylyaparli-Yamatji-Nyungar woman