ChatGPT use and critical-thinking training yield distinct student gains

The experiment at Bocconi University, conducted with OpenAI Economic Research, points to complementary roles for AI and critical-thinking training in education. While ChatGPT access helped students produce more polished and logically coherent work, the causal-reasoning exercise encouraged a wider range of original ideas—an effect that a traditional rubric failed to capture. This gap suggests that as AI makes polished answers easier, evaluations may need to shift to reward originality and reasoning, not just clarity and structure.

The study's randomized design strengthens its contribution to research on AI's impact on learning. Yet the broader challenge for schools remains: assignments may need to evolve to measure qualities like originality, which are not easily quantified. The takeaway is that AI and critical-thinking training are not opposites but complementary tools—each addressing a different aspect of student development. What this means for educators is that integrating both may be key, but the practical path depends on how quickly assessments can be redesigned to value these distinct skills.