A survey reveals that the notion of a "student AI revolt" is largely a myth.

      The boos heard at last spring’s graduations indicated a potential student backlash against AI. However, the data suggests otherwise. At a business school in Washington, the regular use of AI has gradually become commonplace, with the predominant request being for its integration into the curriculum rather than its elimination.

      This insight is based on a three-year survey conducted at American University’s Kogod School of Business, which was shared exclusively with Axios. According to the report, the percentage of students utilizing AI 11 or more times per week rose from 6.2% in 2024 to 29% this year. Additionally, over 80% of students reported using AI for their coursework in the past six months.

      Employers are adapting as well. The occurrence of AI-related questions during interviews increased from 11.6% in 2024 to 42.6% this year, a trend that hasn’t gone unnoticed by students. The skill they most desire to learn is prompt engineering, with coding now ranking second, rising from under 4% to nearly 20% in just one year. This indicates a shift from casual use to preparation for job interviews.

      Usage preferences vary by age group: undergraduates tend to prefer ChatGPT, while graduate students favor Claude. Although Perplexity was the most frequently used tool, all students surveyed had access to free Perplexity Pro through a campus arrangement, which likely inflates its usage figures. When asked to choose a tool, accuracy was the primary concern.

      Despite this, students remain discerning. Almost half of those at Kogod expressed concerns about academic integrity. In a separate survey by Inside Higher Ed, 40% voiced worries that AI might be diminishing their independent thinking abilities. This feedback indicates not a rejection of AI, but a call for clearer guidelines.

      There’s an interesting paradox in the data. Brainstorming continues to be the most common application of AI, a trend sustained over three years. However, researchers at Wharton discovered that individuals generated a broader array of ideas using their own thoughts and online searches compared to when they relied on ChatGPT. The most frequently used application may inadvertently narrow creative thinking.

      Educational institutions are grappling with how to handle AI’s presence. Initially, faculty attempted to create assignments that were “AI-proof,” but this strategy was eventually abandoned, according to Kogod’s interim dean, Casey Evans. The new strategy is more straightforward: staff input a student’s work into AI, inform them that the assessment results in a C, and then help them aim for an A. Resistance to this approach has been minimal.

      In contrast, other universities are taking a different stance. Some have tasked AI with monitoring students during examinations, while others, such as Brown, have reverted to requiring in-person, handwritten papers. Tools for detecting cheating and those for facilitating it are constantly evolving in a cat-and-mouse game. The educational landscape is still searching for a definitive resolution.

      While the Kogod survey numbers have a caveat, being self-reported from 483 students at a single institution, larger surveys echo similar findings: Inside Higher Ed noted that 85% of students had utilized AI for coursework, and Gallup reported a weekly usage rate of 57%. The graduation boos were indeed audible, but so is the adaptation occurring among students. While they may not embrace an AI-influenced job market, they are positioning themselves to thrive within it.