Understanding AI and Learning Outcomes

Current research on AI's impact on learning has mainly focused on short-term outcomes such as test scores. However, there are limits to understanding how students actually use AI in real-world settings and how this affects long-term learning outcomes.

To address this, OpenAI has partnered with the University of Tartu in Estonia and Stanford University's SCALE Initiative to develop the Learning Outcomes Measurement Suite. This framework supports the longitudinal measurement of learning outcomes across diverse educational settings.

OpenAI has also built a new research ecosystem called the Learning Lab. Researchers from Arizona State University, UCL Knowledge Lab, MIT Media Lab, and others are participating, conducting in-depth research into AI's impact on learning.

OpenAI is already validating learning effectiveness through its study mode feature. study mode encourages deep learning through scaffolding, comprehension checks, and guided practice, rather than simply providing answers. A randomized controlled trial involving more than 300 college students produced early results suggesting that this pedagogical approach can improve learning outcomes.