Think outside the bots: how to use AI without your brain turning to mush

GPS once dulled our sense of direction. Search engines weakened our memory. Now scientists warn that AI could compound the problem, threatening everything from creativity and critical thinking to attention span. A few years ago, I started using AI tools every day, partly because I write about them for a living. Recently, though, an emerging crop of studies has me worried. Am I quietly turning my own brain to mush?

The concerns come with a lot of nuance. "On a high level, yes," says Adam Green, a professor of neuroscience at Georgetown University. "If you are not doing as much of the actual thinking, then your capability to do that kind of thinking is going to atrophy." Other researchers worry that students who lean too heavily on chatbots may quietly surrender the cognitive friction that keeps thinking sharp, producing fewer original, novel ideas as a society. The science is brand new, and the answers are far from settled.

Not every expert agrees the danger is real. Twenty years ago, an idea cropped up about digital dementia, with researchers warning that overreliance on technology could damage memory and cognition. A recent meta-analysis of 57 studies covering more than 411,000 adults, co-authored by Jared Benge of Dell Medical School, found no clear evidence for it. In fact, technology use seemed to lower the risk of cognitive impairment. The picture is not all or nothing. Still, related effects do emerge in narrower studies. People who rely on satellite navigation stop building their own spatial memory of the world around them. The Google Effect describes how readers retain less of what they look up, because almost no effort is required.

Don't take AI's word for it. Heavier AI users have scored worse on standard critical thinking tests, apparently because they offload their reasoning onto robots. Researchers at the University of Pennsylvania call this cognitive surrender: trusting AI over your own thinking and intuition, even when AI is wrong. Microsoft Research found the danger sharpens when you know little about a subject. The fix starts before you open the app. Form your own rough view first, then use AI to challenge it. That way the technology is pressure-testing your judgment rather than replacing it.

Add friction to your research. "If something is in front of you and you see it, you often think it is in long-term memory when it is not," says Barbara Oakley, an emeritus professor at the University of Oakland. Some light problem solving before opening a chatbot can sharpen retention later, one unpublished study suggests. When you do reach for AI to remember information, slow down. Take notes by hand if possible. Ask the chatbot to quiz you. It sounds fussy, but the small effort is the point.

Leave the blank page blank a little longer. Research suggests people who use AI for creative tasks produce ideas that are more predictable and less original. Your brain builds creative capacity by reaching across experience, memory, and false starts to make connections only you could make. Hand that work to AI, and you skip the workout. One simple fix: write your own rough notes first, however ugly, before opening any chatbot. Then use AI to refine, develop, or poke holes in what you have got.

Pay attention. The onslaught of technology, including AI, is making focus harder. So apply the same logic everywhere: do things the slow way on purpose. Don't let an AI summarise that long article. Sit with a hard problem before asking a robot. Tech giants are rushing to shove more AI onto our phones, but you can still keep your brain in the driver's seat. Green argues that human thinking is structurally different from AI in ways that really matter: we make connections that are personal, unexpected, and genuinely original, in ways that digital probability machines simply cannot replicate. The desire to think, create, and figure things out for ourselves, it seems, is harder than ever to outsource.