grades

In December 2025, a shooting on the campus of Brown University caused the death of two people. One of them had just introduced herself to one of the professors at that university, the Spanish Robert Serrano. The tragedy made Serrano, an economics professor at said institution, make a decision that would later have a unique ramification: the growing role of AI in academic teaching. Exams from home. That event made Serrano decided to change the evaluation method of its spring 2026 course. The always demanding ECON 1170 would allow take-home exams in both the midterms and the final exam. The subject, with a reputation for being “a bone”, normally attracted few (but extraordinary) students, and Serrano once had classes of 8 students. This spring, upon learning of the new testing method, 86 enrolled. Brilliant students looking for shortcuts. Brown University is part of the prestigious Ivy League, a group of private universities in the northeastern United States. All of them are famous for their theoretical academic excellence and social elitism, and also theoretically those who access it are brilliant students. The thing is that these students tend to be especially ambitious and competitive, and that is where AI comes in: it is an attractive shortcut to overcome the overload of assignments and exams. Too many outstanding. In the midterm exam on March 5, the average grade was absolutely extraordinary: 96 out of 100. In fact, 40 of the 86 students obtained a perfect score of 100 out of 100. These grades were, as Serrano said, surprising because the average grades in previous courses ranged between 65 and 80 points out of 100, “and this exam was more difficult than the ones I took in the past because doing it at home is an opportunity to pose a greater challenge to the class, since you give students unlimited time. Something doesn’t add upto. It wasn’t just that the notes didn’t add up: it was that some of the answers, even though they were correct, were strange. According to Serrano “they had a very convoluted style.” What Serrano then did with his teaching team was try to answer those same questions with ChatGT. There they discovered that the answers were very similar to those they had seen in the exams. In-person exams. With that suspicion already created in Serrano, the professor decided to take the final exam in person. He notified all his students by email: if the distribution of grades in the final was similar to that of the midterm, he would consider the latter valid. Otherwise, it would annul those grades and what would really weigh in the final grade would be the result of the final exam. This notice had an immediate effect: 18 students abandoned the course suddenly, and another nine did not even show up at the end. Of those 27 students who failed the subject, 22 had gotten a perfect 100 on the midterm. The note collapses. Once the remaining students took the final exam, the results spoke for themselves. The average grade plummeted to 48 out of 100, half of what it had been when students were allowed to take the exam from home. The conclusion was clear. The University removes iron. Serrano went blind at age 17 due to retinal dystrophy and still graduated from Harvard. He had no intention of just letting this go, and although he reported it to the Brow administration, the institution’s response was rather lukewarm. Reporting the situation to the media. To make the problem known, he began talking to media outlets such as TheChronicle, The Country either Inside Higher Ed. In this last medium you can see how the students’ grades dropped: of the 59 who appeared, half passed, but only four got 80 out of 100 or more, and only 17 surpassed 60 points. It was then that the academic committee suggested Serrano file an individual complaint for each suspicious student, something he described as “ridiculous.” The dean pointed out that students who cheat “almost never do so from a malicious place,” but rather due to specific pressure. It is not an isolated case. What happened in Brown is not an extraordinary event. The arrival of AI in university classrooms has transformed evaluation methods themselves. in Princeton They just ended a 133-year-old tradition why teachers left the classroom during final exams, since they trusted the honor code signed by the students. A former chair of that Princeton honor committee acknowledged that “there is a feeling that people cheat on homework exams and that people just use ChatGPT,” and the more people believe they cheat, the more they are encouraged to cheat. Image | Magnificent In Xataka | “We exploit the weaknesses of AI”: teachers’ lonely struggle to reinvent homework and exams