Can AI be trusted in schools?

ARTIFICIAL INTELLIGENCE has become a school troublemaker. Not every child will go home and write 800 words on “Macbeth” when ChatGPT can do it for them. In Turkey and the Netherlands, experiments using large language models (LLMs) to teach coding and maths ended with mixed results: some pupils became so dependent on the LLM that, when it was removed, they performed worse than classmates who had never used it. Teachers, too, have learned to cheat. Students complain that some educators are using bots to churn out generic feedback on their work.