AI Brain Fry at Work Is Real, Say Researchers

AI brain fry is now a documented workplace condition — not just a catchy phrase — and researchers say it's costing companies far more than they realize. A study of nearly 1,500 full-time US workers, published in the Harvard Business Review on Friday, found that excessive AI use at work is doing the opposite of what the hype promised: instead of clearing the decks, it's piling more cognitive weight onto already stretched employees.

What Is AI Brain Fry?

What exactly does AI brain fry mean for workers?

AI brain fry refers to mental fatigue that results from excessive use of, interaction with, and oversight of AI tools beyond a worker's cognitive capacity. Researchers from Boston Consulting Group and the University of California coined the term after surveying nearly 1,500 full-time US workers, finding that 14% reported experiencing it.

Affected workers described the condition in visceral terms: a 'mental hangover,' a 'fog' or 'buzzing' feeling, headaches, and an inability to think straight. Slower decision-making and difficulty focusing rounded out the picture. This isn't burnout — it's something distinct, tied specifically to the cognitive load of managing AI tools, not just a general sense of work exhaustion.

Contrary to the promise of having more time to focus on meaningful work, juggling and multitasking can become the definitive features of working with AI.

The Numbers Behind the Fatigue

The data, drawn from the AI brain fry study published in Harvard Business Review, makes the case hard to dismiss. Workers reporting AI brain fry experienced 33% more decision fatigue than those who didn't — a gap researchers say could translate into millions of dollars in annual losses for large employers.

The quit risk is equally striking. Workers experiencing AI brain fry were around 40% more likely to have an active intent to leave their jobs. And on the errors front, affected workers self-reported making nearly 40% more major mistakes — defined as errors with serious consequences for safety, outcomes, or key decisions. Three data points, same story: AI overuse breaks people, not just processes.

Does AI Ever Actually Help?

Yes — when it's used right. Respondents who used AI to cut down on repetitive, routine tasks reported burnout levels 15% lower than workers who didn't use AI that way. That's a real finding, and it suggests the problem isn't AI itself. It's how companies are deploying it.

The distinction matters. Replacing tedious work = reduced burnout. Layering AI oversight on top of an already full workload = cognitive overload. Most enterprise AI rollouts, if we're being honest, tend toward the latter. As multi-agent systems multiply, employees end up toggling between more tools, not fewer — exactly the trap the researchers flagged.

What Do Company Leaders Do With This?

The Coinbase story cuts to the heart of where things went wrong. CEO Brian Armstrong fired engineers who refused to use AI and set a target to have AI write half the platform's code. That's the logical end of treating AI adoption as a performance metric — it incentivizes usage quantity over quality, which is precisely what the researchers warned against.

Their prescription is pointed: leaders should 'clearly define AI's purpose in the organization' and explain how workloads will actually change. Stick to measurable outcomes. Stop counting prompts. 'Incentivizing quantity of use will lead to waste, low-quality work, and unnecessary mental strain,' the researchers wrote. That's not a subtle critique — it's a direct shot at how most companies are currently running their AI programs.