AI adoption among policy professionals is growing, but only slowly. Apolitical’s new report looks into why, and what policy professionals need from AI tools.
Fast tools, high stakes: AI and the policy professional looks at how policy researchers are actually using AI, drawing on interviews, polling and survey data from public servants worldwide. The headline finding is that adoption is growing but uneven, with 44% of policy professionals using AI daily while the rest use it weekly or less. At the same time, the large majority express caution and reservations about using AI in their work.
That caution comes down to three specific concerns: doubts about the accuracy of AI-generated evidence, worry about losing critical thinking skills built over years of practice, and questions about how sensitive information gets handled.
The report argues that this caution should be read as a feature of good policy work rather than a barrier to overcome. Because policy decisions carry public accountability, professionals hold AI to a higher standard for traceability and control than other government functions. This research has shaped how we designed PolicyNova, a purpose-built policy research tool. Its sources are all traceable and trusted, meaning that judgement remains in the hands of the policy professional rather than the tool.