AI-Driven Professional Development for Maths Teachers — Copur-Gencturk's U.S. Randomised Trial of Intelligent-Tutoring Feedback

Copur-Gencturk, Li and Atabas developed a virtual, interactive professional-development programme in which intelligent tutoring systems give teachers just-in-time, personalised feedback on their mathematics content and pedagogical content knowledge.
The programme was tested in a randomised controlled trial with teachers across the United States: 29 were assigned to the programme and 23 to a business-as-usual comparison with no additional training (52 in total). The study was published in the American Educational Research Journal in 2024.
Teachers who completed the programme selected more cognitively demanding tasks and created a more coherent, conceptually focused learning environment than comparison teachers, and reported greater knowledge and confidence in teaching mathematics.
Caveats: there was no significant difference in how the tasks were implemented in class, the sample is small, and no student-achievement outcomes were reported in this study (the author herself calls for larger trials). Scalability is plausible because the programme is delivered virtually, but it has not been shown at scale.

Read the full analysis: https://scale.stanford.edu/publications/improving-teaching-scale-can-ai-be-incorporated-professional-development-create