Summary:
- MIT researchers have identified significant risks regarding the use of Large Language Models (LLMs) in undergraduate education, noting that AI tools can lead to "over-reliance" and the atrophy of critical thinking and foundational problem-solving skills.
- The study highlights a pedagogical paradox where AI can assist in learning complex concepts but simultaneously undermines the cognitive struggle necessary for deep academic mastery, prompting calls for a redesign of assessment methods that prioritize in-class, AI-resistant evaluation.