A mathematics professor at the University of California, Berkeley, has altered his grading method after observing a worrying trend among students. Richard Borcherds, who won the Fields Medal in 1998, believes that the rise of artificial intelligence (AI) has diminished the effectiveness of traditional homework grading. He has expressed doubts about whether students are truly learning from their assignments. While he acknowledges that his new method is not perfect, he feels the issue cannot be ignored.
Borcherds has stopped grading homework based on correctness. He believes that many students are using AI to get correct answers, making it difficult to assess their understanding of the material. In a conversation with The California Post, he noted the widespread use of AI for homework, which can create a strong incentive for students to rely on these tools. He pointed out that grading for correctness might unfairly penalize those who are honest in their work.
The professor is worried that students who work hard to solve problems might score lower than those who use AI to produce polished responses. This means that grades may not reflect the effort and thought put into finding solutions. As a result, Borcherds has moved to a grading system where he gives credit for submitting homework rather than checking each answer for accuracy. He mentioned that this method is not a great solution and that he currently has no better option.
Borcherds also highlighted the difficulty in distinguishing between work done by humans and work generated by AI. He remarked that while AI used to make obvious mistakes, it has become quite skilled and rarely makes errors that are easy to spot. A survey by Common Sense Media conducted in August 2026 showed that 70% of American teenagers reported using AI for school assignments. Among these students, 63% said they got answers directly from AI, whether they used them as they were or modified them.
The survey also indicated that many students recognized the downsides of AI use. Among those who used AI, 38% felt it led them to come up with fewer original ideas, and 39% believed they missed out on important learning opportunities. Additionally, 66% of AI users claimed that the technology helped them understand their coursework better.
However, a research paper from the Center for Economic Policy Research in June 2026 raised concerns about AI’s effects on learning. Researchers analyzed data from 26,811 middle and high school students in China over a period of 30 months. They found that the use of AI was linked to an 18% increase in homework scores and a 30% reduction in time spent on assignments, but there was a 20% drop in monthly exam scores within six months.
The researchers noted that the learning losses were mainly among students who seemed to outsource their homework rather than use AI while maintaining study habits. Given these findings, what are your thoughts on the impact of AI on student learning?
Sources:
Source: Bored Panda, by Lei RV