A March 2026 study published by Harvard Business Review found that certain patterns of AI use can contribute to cognitive fatigue, with workers reporting symptoms including mental fog, difficulty concentrating, slower decision-making and headaches.
The study, conducted by researchers from Boston Consulting Group and other institutions, surveyed 1,488 full-time workers in the United States.
What is AI brain fry?
The researchers describe AI brain fry as mental fatigue that occurs when people use or oversee AI tools beyond what their cognitive capacity can comfortably handle.
The problem is not necessarily asking an AI tool to draft an email, summarise a document or generate ideas.
Instead, much of the strain comes from overseeing AI output, checking whether it is accurate, correcting mistakes, monitoring multiple streams of information and deciding what can be trusted.
Imagine reading a report written by a colleague compared with reading one that you have to fact-check sentence by sentence. Both require attention, but the second demands considerably more mental effort.
AI oversight can create a similar burden.
The numbers behind AI fatigue
About 14% of US knowledge workers already report experiencing this fog-like fatigue.
In marketing departments specifically, that number jumps to 26 percent, likely because marketing roles often involve juggling the most AI tools at once. Workers affected reported more mistakes at work, slower decision-making, and a noticeably higher urge to quit their jobs entirely.
One researcher on the study, Julie Bedard of BCG, put it simply: the AI can move far ahead of us, but people are still working with the same brain they had yesterday.
Why using several AI tools can become exhausting
It seems obvious that more AI tools should mean more help. The data says otherwise.
Productivity actually peaks when someone uses just one to three AI tools regularly. Once a person is juggling four or more tools at once, the mental cost of switching between them starts to outweigh whatever time those tools were supposed to save.
Every extra tool adds another interface to learn, another output to double-check, and another place attention has to jump to.
This lines up with a broader pattern researchers have noticed elsewhere too. The average length of time someone can stay genuinely focused on one task has been shrinking for years, and constant tool-switching appears to be making that shrink even faster.
What AI brain fry feel like
The workers studied described experiences including a ‘buzzing’ sensation, mental fog, difficulty concentrating, slower decision-making and headaches.
The symptoms may not look like an obvious workplace crisis.
Instead, a worker may simply finish a normal day feeling mentally exhausted, struggle to concentrate on routine tasks or find it harder to make decisions.
Over time, that cognitive fatigue can affect how people perform their jobs and how they feel about their work.
How workers can reduce AI fatigue
The answer is not necessarily to stop using AI.
The researchers suggest redesigning the way AI is incorporated into daily work. This can include limiting the number of AI tools employees are expected to manage and avoiding workflows that require workers to constantly monitor multiple streams of AI-generated output.
Workers can also benefit from grouping AI-related tasks into specific periods instead of repeatedly switching between AI tools throughout the day.
Creating periods for human-only work can also give workers time to think, make decisions and complete tasks without constantly evaluating AI output.
Why employers also have a role to play
AI-related fatigue is not simply an individual productivity problem.
The way organisations introduce AI can influence how much cognitive pressure employees experience. Managers can help by clarifying which AI tools workers are expected to use, providing guidance on how to evaluate AI output and avoiding unnecessary duplication of tools.
The researchers also point to the importance of workplace culture and managerial support in reducing the effects of AI-related fatigue.
As AI becomes increasingly embedded in everyday work, the challenge may no longer be simply getting employees to use the technology.
It may be learning how to use AI without overwhelming the people expected to supervise it.