Generative artificial intelligence in policy analysis: a study of UK think tanks

This article explores the use of generative AI (GenAI) in policy analysis as it is conducted by think tank researchers in the United Kingdom.

Think tanks focus their policy analysis activities on the three dimensions of problem recognition and definition, the development of policy ideas, and the influencing of public, politicians and public servants. These three dimensions form the framework for analysis of survey and interview data. For the survey, 554 individuals, identified as think tank ‘policy analysts’ or ‘policy researchers’, from 94 UK-based think tanks were invited to take part. Eight in-depth interviews with individuals from this group were conducted as well.

The data suggests that for UK think tank analysts, GenAI currently plays a rather insignificant role in policy analysis. GenAI is used only in limited ways to help with menial tasks and for dissemination and publication while ideation and policy formulation are still ‘done’ by humans, with only some support from GenAI. It appears also that think tank analysts are aware of the risks of GenAI while they understand and embrace its current opportunities. Further, there is some concern about the lack of concrete policies that govern the use of GenAI at think tanks. The biggest concern revolves around how GenAI could reduce the quality of policy analysis and therefore the quality of policy proposals if used inappropriately.