Participatory AI & Social Justice

Artificial Intelligence (AI) and Machine Learning (ML) technologies are increasingly integrated into a variety of products and services. However, AI promises often fall short and unintended consequences multiply, causing a negative impact, especially on marginalised communities. In response, Participatory AI has emerged as a promising method to account for the consequences on people subject to AI and to help develop systems that are better aligned with societal values. Yet meaningful participation remains difficult to achieve. Participation tends to be short-term and consultative, or even a mask for hidden labour. Moreover, the aim of building knowledge that is situated and subjective often conflicts with the global scale and generalizability of AI, especially when it comes to foundational models. This panel unites experts to explore how to account for these issues, and more specifically, to collectively draw a picture of the limits of Participatory AI in terms of social justice.