AI chatbots may be shrinking the thinking muscle you use every day

The sharpest warning from SXSW London was not about AI replacing jobs or reshaping search. It came from Gloria Mark, a psychologist at the University of California, Irvine, who has spent 30 years studying how people interact with digital technologies. Her message was blunt: when we increasingly let AI tools do the thinking for us, we may be weakening the mental muscles that keep attention, learning, and judgment in shape. In her words, the risk is that our cognitive abilities will weaken over time, because if we are not constantly exercising our muscles, they can atrophy, and, as she put it, that is exactly what can happen with our minds.

That warning lands with more force because Mark is not guessing from the sidelines. She has been tracking how digital habits affect attention for two decades using what she calls living laboratories, where sensors and trackers monitor adult volunteers’ attention, mood, and behavior as they use devices. In 2003, she found that the average user could stay focused on one thing for around two and a half minutes. By 2012, that number had fallen to around 75 seconds. In research she conducted between 2014 and 2020, it dropped again, to a mere 47 seconds on average. That is the backdrop for why she is now alarmed by AI chatbots. If attention was already getting chopped into ever-smaller slices by email, phones, and feeds, AI may be turning down the effort required to think at all.

Mark’s concern is rooted in a simple cognitive idea: effort matters. When people have to evaluate, summarize, or write something themselves, they engage in what she calls depth of processing. That deeper engagement makes people more likely to learn, understand, and retain information. But when users ask AI bots like ChatGPT, Claude, and Gemini to do that work, they are deferring their cognitive work to AI. In her view, that is not just a productivity shortcut, it is a behavioral trade-off with long-term consequences. The immediate upside is obvious. The less obvious cost is that people may stop practicing the kind of active thinking that keeps their minds sharp.

That matters because Mark is also seeing what frequent attention switching does to the body and to performance. She said that in her studies, heart rate monitors showed a direct correlation between switching attention fast and stress going up. Her point is not that technology is inherently harmful, but that the constant mental gear-shifting it encourages comes with a measurable cost. It takes longer to do any single task if attention keeps bouncing around, she said, and that is bad for performance and emotional well-being. For executives, that should ring familiar. Faster tool adoption does not automatically mean faster output if the work itself gets shallower, more fragmented, or more stressful.

The AI warning also sits inside a much bigger debate about digital harm, especially for younger users. A few months ago, Meta, which owns Facebook and Instagram, and Google’s YouTube were ordered to pay millions of dollars in damages to a 20-year-old woman who accused the companies of creating products that led her to develop a childhood addiction. Just a couple of weeks ago, Meta settled another lawsuit brought by a rural school district in Kentucky, which had accused the company of designing addictive products that were harmful to students and had sought more than $60 million to cover mental-health costs. Around 1,200 other school districts are pursuing similar legal action against social media companies. That does not prove social media is uniformly harmful, but it does show that regulators, plaintiffs, and schools are increasingly treating digital design choices as something with real-world developmental consequences.

At the same time, Mark is careful not to flatten the story into a simple tech-is-bad narrative. Social media can create meaningful benefits for some people, including those from marginalized groups. A 2024 survey of LGBTQ+ teenagers found that while some described social media as a place of rejection and fear, others described it as a place of belonging, where they could develop friendships and cultivate their identity. That is a useful reminder for anyone making policy, product, or board-level decisions: the effects of digital tools are not one-note, and the same platform can be both a lifeline and a stressor depending on who is using it and how.

The current scientific verdict on children and social media is also less settled than the loudest takes suggest. Mark said the evidence is to date inconclusive, despite what readers may see in best-selling books on the subject. She is hopeful that large, long-term studies will bring more clarity, and one such effort is underway in Australia, which enacted a social media ban for under-16s at the end of last year. For now, her broader point is not that we should panic, but that we should pay attention to how much effort our tools are doing on our behalf. Her suggested course correction is practical, not Luddite: read a book rather than skimming its summary, meet friends in person when you can, and try not to use GPS where you probably do not need it. As she put it, we love technology and cannot give it up, but we have to learn how to create new life routines. For founders, managers, and investors building or using AI, the big takeaway is uncomfortable but useful: the next productivity wave may also be a discipline test, because the companies that win may be the ones that help people stay smarter while getting faster, not just faster while getting dumber.