There has been quite a bit of research that shows relying on artificial intelligence tools encourages people to stop thinking critically and start deferring to the machine. But you’d be shocked at just how quickly your brain can shut off if you let it. According to a new study by researchers at Carnegie Mellon, MIT, Oxford, and UCLA, just a 10-minute session with an AI assistant can lead to users significantly abandoning their own capacity for reasoning.
To show cognitative offloading in action, the researchers gave two groups of people—one aided by AI assistants and one operating entirely on their own. The participants who were given AI assistants (in this case, a chatbot powered by OpenAI’s GPT-5 model) would have the aid pulled from them without warning during the test, and were left to solve the final three questions on their own.
The study tested two different skills: first, giving a group a set of fraction-based arithmetic problems, and then a set of SAT-style reading comprehension questions. Unsurprisingly, the people using AI tended to solve the math problems at a noticeably higher rate during the AI-assisted portion of the test.
But in those final three questions, where they had their assistant removed, the AI group saw the solve rate fall off a cliff. They had a solve rate about 20% lower than those who had to operate on their own the whole way through. They also had nearly double the skip rate, meaning they simply chose not to solve the questions.
Something similar happened in the reading comprehension test—though the AI-assisted test takers did not see a significantly higher solve rate than those operating without help. Instead, the solve rate was similar until AI was removed, at which point those with AI support available saw a drop off in correct answers and an uptick in skip rate.
Using self-reported AI usage patterns, the researchers noted that people who used AI to get answers showed the largest declines in performance. Participants who used AI for hints rather than to solve the questions for them didn’t see a significant impairment in their performance compared to the control group. But those who decided that AI can do the work for them were largely unable to turn their brains back on when the time came.
The sessions lasted about 10 minutes, suggesting that those who decided to rely heavily on AI to solve problems for them abandoned their critical thinking abilities in a matter of minutes. Given the results were similar across both math and reading-related tasks, the researchers posit that performance decline is a “general consequence of AI-assisted problem solving, not specific to any particular task.”
The findings are also in line with a study Microsoft published last year that looked at cognitive decline among knowledge workers, which found that the more people lean on AI, the worse they perform when asked to work without support. It also echoes a study out of Poland, which found that while doctors are better at spotting cancer risks with AI assistance, they perform worse than the no-AI baseline once that assistance is removed.
Read the full article here
