Key points are not available for this paper at this time.
Generative Artificial Intelligence (AI) systems bring innovative ways of information provision and knowledge delivery. In the public sector, generative AI has the potential to decrease bureaucratic discretion in the decision-making process. Increasing reliance on this technology brings challenges of unfair treatment, colonized responses from the system, and data governance. Because of historical interaction, tribal communities are the most underrepresented in policy planning and implementation. Indigenous communities suffer from the neglect of tribal sovereignty by the U.S. federal government and limited accessibility and literacy in the digital world. Generative AI systems exacerbate these challenges with insufficient tribal input. However, the negative impact can be alleviated with digital equity and knowledge cocreation. Digital equity emphasizes the importance of tribal knowledge representation, and knowledge cocreation focuses on the collaboration between Indigenous communities and relevant actors in data governance for generative AI systems. This study proposes two research questions to discuss tribal knowledge cocreation in generative AI systems: (1) what are the biases in the system responses from the tribal perspective? (2) what are the potential resolutions for these problems? The findings from in-depth interviews with tribal members in the U.S. indicate that the insufficient articulation of tribal culture, the lack of crucial tribal historical events, and the inappropriate appellation of tribal nations are the primary drawbacks in the system responses. From the Indigenous perspective, tribal oral traditions, native publications and documents, and collaboration with tribal governments can address the problems of generative AI responses. This study contributes to the theory development of digital equity and knowledge cocreation in tribal generative AI system responses. Policy recommendations and future research agendas are included in this research.
Wang et al. (Sat,) studied this question.