Education researcher Benjamin Riley says AI is to the mind what a hot dog is to the body: satisfying in the moment, harmful as a regular habit. In an essay published by The Verge on 5 October, he argues that the tech industry misunderstands how human brains work, and that this is why its push into schools is backfiring.
The flaw in “the brain is a computer”
Much of the AI industry treats thinking as input, processing and output. Google DeepMind’s Demis Hassabis has described the brain as a biological approximation of a Turing machine, and Elon Musk has called it “a biological computer.”
Riley leans on the work of Paul Cisek, a neuroscientist at the University of Montreal, to make the opposite case. In Cisek’s model, brains evolved as feedback-control systems. We do not passively receive information and calculate a response. We act on our surroundings to change what we perceive.
Riley’s example is a baseball outfielder. A computer-style model implies the player is subconsciously solving physics equations. In practice, the player simply moves to keep the ball in the same spot in their field of view.
Why learning depends on effort
Humans also built a second feedback loop: culture. Imitation, language, writing and eventually schools allowed knowledge to pass between generations. That process relies on people doing the hard mental work themselves.
Generative AI, Riley argues, offers a constant shortcut around that work. Our appetite for fat made sense when food was scarce and became a liability once it was everywhere. Our instinct to save mental energy works the same way once a chatbot can do the thinking.
The industry is targeting classrooms anyway
OpenAI’s Leah Belsky said in July 2025 that learners make up more than half of ChatGPT’s users, calling it the world’s largest learning platform. Anthropic has since launched Claude for Teachers, offering free premium access to classroom educators.
Riley points to growing evidence pointing the other way. One study from China found thousands of students largely stopped doing homework after adopting AI, with clear damage to their learning. Other research suggests regular chatbot use can make students less critical in unrelated tasks. A recent paper describes the pattern as “cognitive delegation”: the more people hand off thinking, the weaker the habits that support independent reasoning become.
What comes next
Policy is starting to move. Norway has imposed a near-ban on AI for school pupils under 13. A major US teachers’ union has called for chatbots to be kept out of elementary schools, and the Los Angeles and New York City school districts have introduced bans of their own. Riley discloses that he has informally advised Schools Beyond Screens, a group that campaigned for those bans.
His conclusion is not that AI must be avoided entirely. An occasional hot dog does little harm. The risk comes when it becomes the daily diet, and for millions of students, it is heading that way.
Sources: Benjamin Riley, “Our minds aren’t equipped to handle AI,” The Verge (5 October 2026); Paul Cisek, “Resynthesizing behavior through phylogenetic refinement” (2019); Reuters on Norway’s school AI restrictions (June 2026).
