Premio Croce Debate 2026

The rapid diffusion of large language models (LLMs) and generative artificial intelligence into education accelerated dramatically with the public release of OpenAI's ChatGPT in November 2022, followed by competing systems from Google (Gemini), Anthropic (Claude), Meta (LLaMA), and Chinese developers such as Baidu (Ernie) and DeepSeek. Within months, these tools were being used by students for drafting essays, solving problems, and tutoring, prompting initial bans by school systems in New York City, Los Angeles, and several European jurisdictions. Most of those bans were quickly reversed as ministries of education recognized both the inevitability of student use and the potential pedagogical value of the technology, shifting the policy debate from prohibition to integration, assessment redesign, and academic-integrity frameworks. The core dispute centers on a tension between opportunity and risk. Proponents, including UNESCO in its 2023 Guidance for Generative AI in Education and Research and the OECD, argue that LLMs can personalize instruction, narrow learning gaps, support students with disabilities, and relieve teacher workload. Critics point to documented harms: factual hallucinations, algorithmic bias, erosion of critical thinking and writing skills, privacy violations affecting minors, widening digital divides between well-resourced and under-resourced schools, and the unresolved copyright status of training data. Empirical evidence is still thin and contested, with early studies (including work published in Science and by MIT researchers in 2024-2025) suggesting that AI tutoring can boost short-term performance but may impair retention and independent reasoning when used as a substitute rather than a scaffold. Today the regulatory landscape is fragmenting. The European Union's AI Act, which entered into force in 2024, classifies certain educational AI uses as 'high-risk' and imposes transparency and human-oversight obligations. The United States has pursued a lighter-touch approach through the 2023 Executive Order on AI (partially rescinded in 2025) and Department of Education guidance, while China has issued strict generative-AI rules through the Cyberspace Administration and is simultaneously promoting AI literacy in its national curriculum. UNESCO has called for a minimum user age of 13, mandatory teacher training, and protection of human agency in learning. Open questions before deliberative bodies include how to assess student work authentically, how to ensure equitable access, how to protect minors' data, and how to preserve the formative cognitive labor that education exists to cultivate.