Learning theory now has to account for the machine in the room
The accessible claim is not “let AI teach everything.” It is that schools need a theory for how students learn with a generative partner beside them. That means teaching students how to question outputs, share agency, notice offloading, and keep thinking about their own thinking.
What the source reports
The public arXiv abstract says older learning theories have limits as generative AI spreads in schools. It names four older theories. The authors say those theories came before AI systems could generate and reason about knowledge. The article proposes Generativism as a learning theory for the generative-AI age. It says learning now happens through back-and-forth knowledge building between people and AI systems. The framework uses four constructs about partnership, agency, AI literacy, and thinking about thinking. The claim is plain: students learn with the tool, not just from it. This is not a test score. It is a lens for class work now. The abstract presents this as a base for rethinking design, learning, assessment, and expertise when generative AI is part of cognition.
Original source
- Title
- Generativism: Toward a Learning Theory for the Age of Generative AI
- Author
- Li, Shan & Zheng, Juan
- Publication
- arXiv
- Date
- Monday, May 18, 2026