In this report, the role of artificial intelligence (AI) in education is analysed using a science-based account of human learning. The report challenges the notion that AI will alter what is important for students to learn or the way they learn from a cognitive perspective.

In this report, the potential for generative large language models to contribute to education, as well as its associated risks, are discussed. Particular attention is paid to human cognitive architecture and its implications for teaching and learning. The constraints these considerations place on the appropriate use of large language models (and indeed any technology) in education are explored. As we will see, from a cognitive perspective, it will remain important for young people to learn to read and write, and to become numerate. Despite the existence of AI technology and online resources, it will also remain important for them to have knowledge and to grasp concepts.

Drawing on scientific research on human learning, the report gives guidance on how AI can help students learn better, and how to avoid potential pitfalls. The author argues that educators must understand appropriate ways to use AI with their students.