Envisioning Communities: A Participatory Approach Towards AI for Social Good.

Bondi E, Xu L, Acosta-Navas D, Killian JA. Envisioning Communities: A Participatory Approach Towards AI for Social Good. AIES '21: Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society. 2021; Pages 425-436. https://doi.org/10.1145/3461702.3462612

Summary: This paper critiques the utilitarian approaches often used in AI for social good (AI4SG) research, arguing that they can marginalize underrepresented communities by prioritizing the majority. The authors advocate for using the capabilities approach, which emphasizes equity in human welfare, to assess AI4SG projects. They introduce the PACT framework, which involves the active participation of affected community members in defining and guiding AI projects to ensure the outcomes align with local needs. The paper offers an incomplete set of guiding questions for conducting participatory AI research that respects a community's unique vision of social good.