HardwareFreaks

Tag: Higher education

  • Professor suspects AI cheating makes finals in person only 2 scores match midterms

    Featured image Professor suspects AI cheating makes finals in person only 2 scores match midterms

    The pursuit of academic integrity often comes with a dose of academic paranoia, and this is perhaps nowhere more evident than in the world of higher education where the lines between effort and assistance are constantly being redrawn. For Brown University professor Roberto Serrano, a recent assessment of student performance took an unexpected turn from grading to suspicion.

    When it came time for a take-home midterm exam in his Welfare Economics and Social Choice Theory class, Professor Serrano observed results that seemed almost too perfect. The class, which usually numbers around 30 students, swelled to 86 after the initial assessment. While historically the course saw midterms averaging between 65 and 80 percent, this semester’s class scored an impressive 96 percent. Initially, some might chalk this up to a particularly gifted cohort, but Serrano sensed something deeper.

    The shift in performance immediately raised eyebrows. The professor noted that while the answers were technically correct, they possessed a style that felt convoluted and strangely polished. This led him to suspect external assistance, specifically the rise of artificial intelligence. Running a comparison with tools like ChatGPT confirmed his growing suspicion: many submissions displayed patterns strongly indicative of AI generation.

    Driven by this suspicion, Serrano took a decisive step. Rather than accepting the results as legitimate, he decided to make the final exam in-person, hoping to eliminate the possibility of cheating entirely. This move triggered further upheaval; eighteen students dropped the class after the announcement, and nine others skipped the final altogether.

    The resulting data was stark. Out of the remaining fifty-nine students, the quality control was severely tested. Only two students received a final grade within ten percent of their midterm score, and only one managed to surpass their prior performance on the final. Three students scored zero.

    This episode highlighted a broader tension between technological advancement and educational standards. Professor Serrano raised his concerns with the university’s Standing Committee on the Academic Code, but it seemed that action was not taken until the story began circulating widely. Now, while the institution is reportedly reviewing each case individually, the professor remains deeply worried about the future of intellectual honesty.

    “We cannot afford to have a society in which a significant fraction of our best young minds think that cheating is OK,” Serrano reflected. “That leads to a declining society, to a failed society. We cannot choose to become idiots.”

    Buy on Amazon

    Credit: Tom’s Hardware

  • Professor’s chart exposes AI cheating scale

    When an academic seeks to understand the modern learning environment, sometimes the most profound insights come not from theory, but from a simple, stark comparison. This is exactly what happened when a professor at Brown University stumbled upon a startling truth about how generative artificial intelligence is reshaping the world of higher education.

    The situation began with Robert Serrano, an economics professor at Brown University, who harbored a suspicion that quickly escalated into a quantifiable reality: most of his students had likely used AI tools to cheat on a take-home midterm exam. It was a classic academic dilemma, but in this instance, it became a flashpoint for a broader discussion about intellectual integrity in the age of algorithms.

    Instead of simply relying on intuition, Professor Serrano decided to run an experiment that cut straight to the heart of the matter. He decided to compare the scores from the take-home midterm with those achieved during an in-person final examination. The resulting data was anything but subtle; it offered one of the clearest demonstrations yet of generative AI’s influence on academic performance.

    The comparison wasn’t just about grades; it was about assessing how learning truly occurred and whether the process of taking a test could be easily automated by machine intelligence. The contrast between performance in an unsupervised setting (the midterm) and a supervised, in-person environment (the final) laid bare the complex relationship between human effort, genuine comprehension, and artificial assistance.

    This unexpected correlation served as a powerful illustration for the academic world. It moved the conversation beyond simple accusations of cheating and into a critical examination of what it means to learn and assess knowledge today. The finding suggested that AI was not just a tool for shortcutting answers but a disruptive force fundamentally altering the educational landscape.

    Professor Serrano’s exercise quickly became more than an internal university matter; it became a powerful, tangible example of how generative AI is impacting academic standards and the very definitions of assessment in academia. It provided a timely and compelling case study on navigating the shifting boundaries between human ingenuity and machine capability in the classroom.

    Credit: TechSpot