- MIT, Oxford and UC Berkeley researchers found that just 10 minutes of AI assistance significantly impairs focus and persistence on difficult tasks.
- Study of 1,222 participants showed those relying on AI quickly abandoned problems when the tool was removed.
- AI users showed weaker neural activity compared to those who thought independently, according to separate MIT Media Lab research.
- Experts warn AI use causes "cognitive atrophy" and may create a generation dependent on algorithmic thinking.
- Researchers urge developing foundational critical thinking skills before introducing AI tools in education.
Brain rot accelerates: Research quantifies AI's cognitive toll
A landmark study presented this month at the Conference on Language Modeling has confirmed what many suspected: reliance on artificial intelligence tools rapidly diminishes human cognitive abilities. Researchers from MIT, Oxford University, Carnegie Mellon, UC Berkeley and UCLA found that spending just 10 minutes using AI assistance on math and reasoning tasks significantly impaired participants' ability to focus and persist with difficult problems. The findings arrive after "brain rot" was named Oxford's 2024 word of the year, reflecting a 230 percent increase in usage as digital technology's toll on cognition entered public consciousness.
The 10-minute threshold
The research team conducted three randomized controlled trials involving 1,222 participants. In the first experiment, 354 people were divided into two groups tasked with solving 15 fraction problems. One group worked independently while the other could consult ChatGPT. The AI-assisted group initially showed higher accuracy. After 12 problems, researchers removed the AI tool. The group that had relied on AI immediately began answering incorrectly or giving up entirely, while the independent group persisted and achieved greater success.
A second experiment with 667 participants produced identical results. A third trial using SAT reading comprehension questions with 201 participants confirmed the pattern: AI-dependent participants cratered in both persistence and accuracy when the crutch was removed.
Historical context for a cognitive crisis
This rapid cognitive degradation marks an unprecedented acceleration in technology's impact on human thinking. Author and computer scientist Cal Newport noted that smartphone-related cognitive declines took years to emerge in data. With AI chatbots, the negative impact appears immediate. The finding carries historical weight: previous transformative technologies—the printing press, radio and internet—expanded information access while leaving analysis and synthesis to human minds. AI uniquely performs those higher-order functions, effectively allowing users to outsource thinking itself.
The "productive struggle" lost
UC Berkeley research fellow Brian Christian, a study co-author, described the work as revealing how AI eliminates "productive struggle"—the difficult mental work that builds cognitive strength and helps individuals discover their intellectual passions. The researchers warned that when AI produces answers in seconds, users become habituated to that speed, making slower, effortful thinking feel intolerable.
Education at a crossroads
Harvard education experts expressed particular concern for developing minds. Tina Grotzer, a principal research scientist at Harvard's Graduate School of Education, emphasized that human cognition exceeds purely statistical AI processing. She noted research showing kindergarteners used strategic information more effectively than Bayesian approaches, highlighting capabilities unique to human minds.
The findings challenge policymakers pushing AI integration into classrooms. Harvard senior lecturer Dan Levy cautioned that students using AI to produce outputs rather than engage in learning miss the fundamental purpose of education. Christopher Dede, a senior research fellow, warned that AI should function as a tool for growth, not a crutch that replaces thinking.
The choice ahead
The research presents a clear verdict: unchecked AI reliance does not augment human intelligence—it replaces it with algorithmic approximation. Study author Brian Christian called for developing AI systems that default to instructive, tutorial-style responses rather than providing instant answers. The technology's future depends not on its capabilities but on whether users will demand tools that build rather than bypass their cognitive capacities. As one research scholar philosopher cautioned, taking shortcuts without knowing the map leaves travelers lost when their navigator fails. The question now is whether society will choose to cultivate the maps or surrender to the convenience of being told where to go.
Sources for this article include:
Please contact us for more information.