The Massachusetts Institute of Technology (MIT) is rethinking how it teaches and evaluates students as artificial intelligence (AI) becomes more influential in higher education. A report by a committee of MIT professors and students found that AI tools are making some traditional academic tasks, like assignments and take-home exams, less effective at measuring students' true understanding. The report also warned that overreliance on AI could weaken students' confidence in their own abilities. To address this, MIT is exploring alternatives such as oral exams, long-term projects, and student portfolios to better assess learning. According to a study from September 2024, students who used AI for their work were less productive and had a lower chance of future success. The data showed that students who performed poorly in their classes were more likely to use AI to make up for their skill gaps. Nearly 15 percent of younger students and 53 percent of older adolescents used AI tools for their schoolwork. While the use of AI is becoming more common in education, the report emphasized the need to differentiate between cheating and using AI as a study aid. MIT wants to increase human interaction in the classroom, as it has noticed students are turning more to AI for help with their studies. The committee spent five months analyzing how MIT students use AI and its effects on their learning. It found that study groups and visits to professors during office hours are becoming less frequent. Traditional assignments and exams are also less effective in evaluating students’ knowledge, as AI can now generate responses to complex university-level questions. This makes it harder for teachers to determine if students truly understand the material. The report also introduced the term "cognitive surrender," describing how some students might use AI to avoid thinking through difficult problems on their own. MIT students frequently use AI, with 46 percent of undergraduates reporting daily use of large language models (LLMs), which are advanced AI systems that can generate text. However, 90 percent of these students expressed concerns about becoming too dependent on AI. The report noted that students are increasingly worried that AI might make them replaceable rather than more capable. MIT is not planning to monitor students’ AI use closely but is instead focusing on changing the way students are assessed. The committee recommends more collaboration in person, as well as a reevaluation of the importance placed on grades and GPA. It also warned against replacing human research assistants with AI solely for cost savings. In February 2026, a study by Epson found that 58 percent of teachers believed students using AI to complete homework were harming their learning. In France, 74 percent of teachers felt traditional resources were essential for learning the basics of reading, writing, and arithmetic. Many also linked a decline in exam results to the use of AI, as students struggled to perform well without it. While teachers did not oppose technology, they stressed the importance of mastering foundational skills first. In New York City, public schools are banning the use of generative AI for over half a million students in the 2026–2027 school year. One strict professor stated, "I receive the same salary whether I pass you or fail you," and has begun failing students who use AI to complete their assignments.