Public say versus public sway: ‘Lived experience literacy’ in the pursuit of participatory AI

Artificial intelligence (AI) technologies are increasingly embedded and promoted across many parts of society. At the same time, concerns persist about their potential to propagate social harms, especially for marginalised and minoritised communities. These concerns include the risk of disconnection between those designing, developing and deploying AI systems and those most affected by it – with the interests of the latter then going underserved or being undermined. Efforts to address these challenges include calls to improve critical AI literacy among the general population, and a drive towards ‘participatory AI’, which seeks to actively involve those most affected by AI technologies in their development. The former is often said to be a necessary precursor to the latter. Nonetheless, while participatory AI's popularity is growing, its implementation does not always uphold its ideals. Participatory approaches executed poorly risk becoming tokenistic or extractive, undermining their intentions. In this commentary we argue that there is an equally important complement to AI literacy: what we call ‘lived experience literacy’ among AI technologists. Lived experience literacy refers to the capacity to recognise and engage with lived experience as a legitimate form of expertise, and to design participatory processes that value and act on that expertise rather than merely soliciting it. Lived experience literacy, which includes needing a paradigm shift in the value afforded to expertise gained through first-hand experience, can help move participatory AI from rhetoric towards practice, mitigating the risks of ‘participation-washing’, and laying groundwork for more meaningfully participatory forms of AI development.