IA generative and critical thinking: it is not the use that counts, but the way to share the work and rework the answers

A survey of 891 American and Turkish students shows that AI's experience only develops critical thinking if tasks are distributed with the tool and their productions are reprocessed.

Does generative AI hinder students' critical thinking, or does it help them? Rather than measuring simple exposure to the tool, one team studied how users organize their collaboration with it, applying the theory of transactive memory systems, originally designed for human teams, to the human-AI duo. The survey involved 891 students in the United States and Turkey.

Two decisive behaviours

Students' confidence in their own thinking abilities, like their confidence in partnering with AI, influences the amount of experience accumulated with the tool, but neither directly predicts the level of critical thinking. The experience itself only has an effect through two practices: distributing and coordinating intellectual work with AI, and actively reworking its responses instead of repeating them as they are.

This structure is found for the most part in both countries, with two exceptions that the authors consider interpretable. For teachers, the message is clear: learning how to redesign AI productions is more important than whether or not to use it.