Student AI Usage: 84% Now Rely on Tools for Schoolwork

84% of high school students used artificial intelligence for schoolwork between January and May 2025. The College Board reported this figure in a study released last fall. It represented a five percentage point jump from the 79% recorded at the start of that period. Approximately half of these students used the technology to brainstorm ideas, revise essays, or research topics.

The rise in usage has forced educators to confront a long-standing issue with student engagement. Alex Campbell, a sociology teacher at Elizabethton High School in Tennessee, observed that generative AI has removed the excuse for schools to avoid addressing low motivation. He argued that the technology did not create the problem. Instead, it exposed the fact that many students find standard assignments uninteresting.

Campbell’s own classroom experience illustrates the alternative approach. In 2017, he tasked his students with solving a series of unsolved murders known as the Redhead Murders. The victims were at least six young women killed between the late 1970s and early 1990s. His students tracked an FBI profiler, interviewed family members, and built timelines from old case files. They identified new victims and developed a theory supported by the Tennessee Bureau of Investigation. Their work became the basis for a 2024 true-crime podcast and an Amazon Prime Video series that entered the top three rankings in July.

Campbell told The 74 that the project succeeded because it was difficult. He said, “I make it obvious to them up front that this is going to be very hard, but they’re going to be invested, and so that makes it easy to do hard work.” He noted that meaningful life experiences are almost always hard. This concept aligns with the idea of “hard fun,” a term popularized in the mid-1980s by MIT Media Lab co-founder Seymour Papert. Papert wrote that students often find difficult coding assignments fun specifically because they are challenging.

Researchers warn that AI is changing the social contract between schools and students. Chris Hulleman, a University of Virginia researcher who founded the university’s Motivate Lab, said students now question the purpose of their assignments. If the work seems boring, they use free AI tools to complete it quickly. A separate University of Southern California report found that most college students default to using AI for “executive help” rather than “instrumental help.” This means they seek fast answers with minimal effort instead of using the tool to deepen understanding.

Colorado math teacher Dylan Kane compared AI to indoor plumbing. Just as plumbing freed people from hauling water but reduced their daily exercise, AI automates cognitive effort. Kane suggested that chatbots could reverse the Flynn effect, a decades-long rise in IQ scores attributed to a cognitively demanding world. He stated, “People are getting dumber, and this is a cross-generational phenomenon.”

Rebecca Winthrop, a senior fellow at the Brookings Institution, noted that AI arrived during a time when school culture was already transactional. A 2024 Gallup survey found that just 17% of middle and high school students strongly agreed that they learned something interesting at school in the previous seven days. Fewer than half said their coursework challenged them. Winthrop argued that the bar for education has shifted higher, requiring more advanced thinking skills than basic literacy and numeracy.

Eric Anderman, an educational psychologist at Ohio State University, said students typically view school as a place to get grades rather than to learn. When teachers emphasize tests, cheating rises. When the focus shifts to learning, dishonesty falls. Anderman often tells pre-service teachers, “‘A’ stands for ‘artificial.’ We made it up.”