Published: June 05, 2026, 13:20 UTC
In 2003, Gloria Mark, a psychologist at the University of California, Irvine, began measuring how long people could focus on a single task while using digital devices. She set up “living laboratories,” equipping adult volunteers with sensors and trackers to monitor their attention, mood, and behavior throughout the workday.
The average attention span she found was two and a half minutes. It surprised even her.
“That surprised me at the time,” Mark recalled during a session at SXSW London this week. “I thought: wow, this is really short.”
She repeated the experiment in 2012. The average had fallen to 75 seconds. By the time she compiled data from 2014 through 2020, the average attention span had dropped to 47 seconds.
Mark’s work, spanning three decades of tracking how people interact with screens, was the subject of a feature interview published June 5 by MIT Technology Review, written by senior reporter Jessica Hamzelou. The headline distilled her concern into a sobering question: “Are AI chatbots making us lose control of our brains?”
The Stress of Switching
Mark’s findings go beyond the headline numbers. She also measured what happens physiologically when people switch their attention frequently. Subjects wore heart rate monitors throughout the day. The correlation was direct: more frequent switches correlated with elevated heart rate and higher stress levels.
“There’s a direct correlation between switching attention fast and stress going up,” Mark said.
The productivity cost is equally clear. Each switch carries a cognitive overhead: the brain must disengage from one task, reorient to a new context, and re-engage. The more switches, the longer any single task takes to complete. Frequent switching degrades performance and emotional well-being simultaneously.
“It just takes longer to do any single task if you are switching your attention,” she said. “It is not great for performance. It is not great for our emotional well-being.”
The New Variable
Mark’s early work focused on email, web browsing, and desktop applications. The landscape has changed dramatically. AI chatbots have become deeply integrated into daily life in just a few years, used for writing, research, problem-solving, creative work, and casual conversation.
Mark told MIT Technology Review she is worried about the impact of AI on attention. The concern is not that chatbots are inherently distracting, but that they make it even easier to offload cognitive work, reducing the need for sustained focus and deep thinking. When an AI can summarize an article, draft an email, or generate code in seconds, the incentive to maintain focused attention weakens further.
The broader worry, shared by other researchers in the field, is that the same cognitive muscles that support sustained attention, critical thinking, and metacognition may weaken with disuse, a phenomenon sometimes called “cognitive debt.”
What About Children?
Mark’s research focused on adults, but the question of how digital technologies affect children has taken on new urgency. In recent months, Meta and Google’s YouTube were ordered to pay millions in damages in a case brought by a young woman who accused the companies of creating addictive products that led to childhood addiction. Meta settled another lawsuit brought by a school district in Kentucky, which had sought more than $60 million. Around 1,200 other school districts are pursuing similar legal action.
Mark cautioned that the evidence on social media’s effects on children remains inconclusive, despite strong opinions on both sides. Large, long-term studies are needed, she said. An effort of this kind is underway in Australia, which enacted a social media ban for under-16s at the end of 2025.
What Can Be Done
Mark emphasized that awareness of the problem is the first step. Recognizing that frequent switching comes at a cognitive cost can help individuals make deliberate choices about how they use technology. Structuring work into focused blocks, reducing unnecessary notifications, and setting boundaries around AI tool use are practical strategies.
The challenge is that these are individual solutions to a systemic problem. The technologies are designed to capture and hold attention; the economic incentives of the attention economy run directly counter to the cognitive habits that support deep focus.
Mark’s data shows that whatever the underlying cause, the trajectory has been consistent for twenty years. Whether AI accelerates that trajectory or provides tools to reverse it is an open question, and one that will likely define the next phase of the human-computer interaction research agenda.
Reference
Hamzelou, J. “Are AI Chatbots Making Us Lose Control of Our Brains?” MIT Technology Review, June 5, 2026. https://www.technologyreview.com/2026/06/05/1138427/
Mark, G. Attention Span: A Groundbreaking Way to Restore Balance, Happiness and Productivity. Hanover Square Press (2023).