https://www.youtube.com/watch?v=1cz0sUKYJ2Y&t=615s

Public Voices in AI is a research project that aims to ensure that public voices are front and centre in artificial intelligence research, development and policy (RD&P). This twelve-month project draws to a close at the end of March 25, and has been synthesising, reviewing, building and sharing knowledge about public views on AI, as well as engaging diverse publics in AI RD&P, with and in consultation with target beneficiaries, including: The UKRI RAI UK programme and affiliates; The broader RAI community; The AI RD&P community not yet committed to responsible AI; and Members of the public, especially from groups most negatively affected by and underrepresented in AI RD&P.

This paper will present initial findings from an evidence review of work on public views of and engagement in AI. It will share findings, such as identified gaps, and joining up existing evidence bases. Evidence reviews tend to synthesise headline findings from research, the favoured method being the ‘systematic review,’ which involves searching databases using keywords, generating a list of relevant research, and assessing the quality of this research through various criteria. This approach validates certain knowledges and invalidates others, omitting some publications (Martín-Martín et al., 2019), legitimising knowledge which has already been legitimated, and excluding knowledge from certain actors (Oman and Bull 2022). The evidence review will go beyond synthesising what other research states are public views of AI, and will extend method to exceed the systematic review. As well as searching databases of academic, grey and practice-based literatures, a ‘snowball’ sampling approach will be taken to locate the sources of evidence that informs the literatures. An analysis framework, similar to that used in the Living With Data Evidence Review (Kennedy et al 2020) will be extended to assess the quality of methods behind the reviewed evidence (eg survey questions, sample size, participant recruitment, motivations for research and its financing) and include criteria that centre equity and inclusion.