Rethinking evidence-informed policy and practice in the age of generative artificial intelligence

The rise of generative artificial intelligence tools such as ChatGPT is transforming how knowledge is produced, accessed and used across educational systems. While these technologies offer new opportunities for efficiency and scalability, they also challenge foundational assumptions about what counts as evidence, how it is interpreted and who holds epistemic authority. The article argues that models of evidence-informed policy and practice (EIPP), although still valuable, require recalibration to address the emerging demands of educational contexts mediated by artificial intelligence. Drawing on recent research in artificial intelligence, critical thinking and professional practice, the article proposes a reframing model – EIPP-CT (Evidence-Informed Policy and Practice with Critical Thinking) – that places critical thinking at the heart of evidence use. Here, critical thinking is conceptualised not merely as a cognitive skill, but as a professional stance encompassing interpretive judgement, epistemic reflexivity and ethical responsibility. The article outlines key risks of over-reliance on content generated by artificial intelligence, including automation bias and diminished transparency, and advocates for institutional safeguards and professional development that foster context-sensitive, deliberative engagement with evidence. It concludes by calling for more systematic research and policy attention to the shifting epistemic landscape of education. In doing so, it aims to preserve the integrity of evidence-informed decision-making in a world increasingly shaped by algorithmic technologies.