Homework quality and independent mastery
Students produced ambitious AI-assisted work, but an unaided exam exposed weak performance. The course did not reliably demonstrate its foundational learning goals.
Why it matters for leaders
Pair assisted projects with meaningful independent checks and adequate unaided practice. Align grading incentives so those checks measure effort and mastery rather than strategic non-participation.
The evidence
Forty-two active coursework records; 27 exam takers averaged 20.6 of 80 points. Evidence also included a student survey and staff observations.
Who was studied
University physics, Germany: indirect K-12 implications
What the evidence cannot tell us
There was no randomized comparison. Grading rewarded homework heavily, reducing incentives to attempt the exam; course difficulty and design also confound the results. This cannot isolate an AI effect.
Read the original sources
Summary and leadership analysis from the August 31–September 6, 2026 brief. The edition’s selection notes explain source access and evidence limitations.