Assessing AI to make sure it is ethical and responsible has, until now, been left solely in the hands of AI experts. There is no agreement of what counts as ‘good enough’ to be acceptable to stakeholders, including ones who ultimately use the AI or are affected by its outputs. PHAWM, one of three RAI UK-funded keystone projects, is driving change in AI testing and evaluation through the novel concept of participatory AI auditing, where a diverse set of stakeholders without a technical background in AI, such as domain experts, regulators, decision subjects and end-users, are empowered to undertake audits of predictive and generative AI, either individually or collectively. This talk will describe the landscape of AI assessment, particularly from a stakeholder perspective. [More Info]