Richard Borcherds, a UC Berkeley prof and Fields medalist, no longer grades his students’ homework.
After seeing how many of his students used AI to consistently obtain top scores, he felt there was “no point” in paying readers to class for correctness in homework, as AI use is now unavoidable.
Borcherds revealed this in a sit-down discussion Tuesday, following OpenAI’s notice that AI agents apparently solved one of mathematics’ most difficult unsolved problems, the Navier-Stokes beingness and smoothness problem.
“If AI continues to enhance at current rate, then, in a few years, there’ll be nothing individual mathematicians can do that AI can’t do better,” Borcherds said.
Borcherds stated he currently grades on completion if a pupil hands in their homework. He admitted that it is “not a extremely fine system,” but at the moment, he has no alternative.
Borcherds stated he no longer knows whether to advance students to go into the math department, as it seemed five years ago that being a mathematician was among the safest careers from the AI boom.
“I have no idea what to advise. I equal asked ChatGPT this exact question, and equal ChatGPT stated it didn’t have any idea what was going to happen,” Borcherds said.
He described a plentiful amount of “inertia” in the scholarly scheme and stated he wouldn’t be amazed if things stayed the identical in the math division for a brace years notwithstanding AI developments.
He drew a difference to bodily document journals, which publish equal although they have been “utterly pointless” for two decades.
“There have been plentifulness of specialized breakthroughs in the former — hand weaving was put out of endeavor by looms, and horses were put out of endeavor by motor cars, and so on,” Borcherds said. “This is the archetypal period item akin this has happened in mathematics, though.”
The Berkeley Division of the Academic Senate asks that department contain a declaration concerning way expectations including the use of AI tools for coursework in way syllabi, according to an email from Anne Marie Richard, assistant vice provost for undergraduate learning and chief scholarly innovation officer. However, she stated the straightforward policies on AI use can change dramatically alongside way flat and concentration area.
Campus has tried multiple AI finding tools, including Turnitin and Rumi, but have resolute that these tools aren’t adequate adequate to be standardized on campus.
AI guideline for classes is up to the discretion of faculty, but there is a expanding repository of example AI use statements accessible in the AI Hub, the evidence directory for AI tools, training and policies, for department to use in their syllabi, according to Richard.
