Research from Boston Consulting Group and the University of California, Riverside surveyed nearly 1,500 workers and found that managing too many AI tools at once can push employees past their cognitive limits, a state the authors term “AI brain fry.”
AI is supposed to give workers their brains back. Delegate routine tasks to automated tools, and employees have more headspace for strategic thinking and complex problem-solving. A study published in the Harvard Business Review in March 2026 suggests the reality is more complicated. Researchers surveyed 1,488 full-time US workers and found that overseeing too many AI tools simultaneously can push employees beyond their cognitive limits. The authors define the resulting condition — which they call “AI brain fry” — as mental fatigue from excessive use or oversight of AI tools beyond one’s cognitive capacity.
Mental fog
In interviews, affected workers described symptoms including mental fog, difficulty concentrating, slower decision-making, compulsive double-checking, and, in some cases, physical headaches that prompted them to step away from their screens. The researchers are careful to distinguish brain fry from burnout. While burnout tends to affect physical and emotional wellbeing over an extended period, brain fry specifically targets attentional, working memory, and executive systems — the cognitive resources the brain draws on when monitoring and directing AI outputs.
One senior engineering manager described the experience: “I had a dozen browser tabs open in my head, all fighting for attention. What finally snapped me out of it was realising I was working harder to manage the tools than to actually solve the problem.” Workers with high AI oversight demands — those actively reading, interpreting, and directing what AI agents produce — reported expending 14 percent more mental effort at work, feeling 12 percent more mentally fatigued, and experiencing 19 percent greater information overload compared with colleagues who had lower oversight responsibilities.
The cognitive load translated into measurable business outcomes. Workers experiencing brain fry logged 39 percent more major errors, reported a 33 percent rise in decision fatigue, and showed roughly a 10 percent higher intention to leave their jobs. The research also identified a productivity curve. Moving from one AI tool to two produced measurable gains. Three tools represented roughly the peak. Beyond four, productivity began to decline — running into the same ceiling that conventional multitasking research has long documented.
Marketing and HR at risk
Brain fry is not evenly distributed across roles. Marketers reported the highest prevalence at 26 percent, followed by HR professionals at 19 percent and software engineers or IT workers at 18 percent. Legal teams reported the lowest rate at 6 percent, which the authors suggest may reflect the precision demands of legal work that have so far limited wholesale AI delegation in that field. Brain fry appeared disproportionately among workers identified as high performers — those who had leaned most fully into AI adoption.
The researchers draw a distinction that shapes the study’s practical implications. Using AI to automate repetitive, routine tasks does not appear to cause brain fry — in fact, it can reduce burnout by removing drudgery from the workday. The risk lies specifically in oversight: the human-in-the-loop model where workers are responsible for reviewing, correcting, and directing what AI agents produce.
The authors frame the problem as an organisational challenge rather than an individual one. Their recommendations include limiting the number of AI agents any one worker oversees at a given time, building AI outputs into automated workflows rather than requiring manual processing, revisiting performance metrics that reward volume of AI use over quality of outcomes, giving staff adequate time to develop AI skills, and ensuring leadership actively addresses cognitive concerns raised by employees.
