Indigenous Data Rights As a Tool to Interrogate and Counter Algorithmic Bias

As Artificial Intelligence (AI) systems become increasingly pervasive, the risks of algorithmic bias, discrimination, and data exploitation continue to disproportionately affect Indigenous peoples. In addition to amplifying existing data harms, AI systems provide new ways to sort, profile, exclude, exploit, and discriminate. Further, the complexity, opacity, and proprietary nature of AI systems make it difficult to access crucial information to understand the system and its decision making processes. Indigenous Data Sovereignty (IDSov) offers a framework to address these concerns by empowering Indigenous communities to assert control over their data and and advocate for transparency and accountability in AI systems.

This presentation will explore how IDSov principles, such as the First Nations Principles of OCAP® (Ownership, Control, Access, and Possession), can be applied to interrogate AI systems, particularly in terms of transparency regarding the training sets used and the input data accessible to these systems. While it may be impossible to eliminate bias from AI entirely, the application of Indigenous data rights can play a crucial role in surfacing and mitigating its harmful effects. IDSov provides tools to challenge and improve the design and functioning of AI technologies for Indigenous users and data subjects. Similar to efforts in Explainable AI (XAI), which aims to make AI decision-making processes more transparent and understandable, Indigenous Data Sovereignty advocates for greater accountability in how AI systems are built and deployed. By asserting ownership and control over their data, Indigenous communities can demand insight into the data sources and algorithmic structures of AI systems, scrutinizing the extent to which they reflect or exacerbate existing inequities. Indigenous Data Sovereignty is not only a protective measure but a proactive tool for creating more just and inclusive AI systems.