The Office Blueprint: ‘AI brain fry’ a real risk for office workers in the push to productivity
Monday, 23 March 2026
One of the most popular X posts in recent times made a point that an increasing number of people can relate to.
“I end each day exhausted …six worktrees open, four half-written features, two ‘quick fixes’ that spawned rabbit holes, and a growing sense that I’m losing the plot entirely,” wrote engineer Francesco Bonacci, the founder of Cua AI.
Bonacci is up to his proverbials in AI tools and platforms, and so overdosing on AI is part of his job description. But increasingly, overdosing on AI is what office workers risk with a tsunami of AI tools washing over the workplace, aiming to make us quicker, more efficient, and “able to spend time” on what we love ‒ or so we’re told.
This includes New Zealand workers, even if we are constantly told we are trailing the pack in our AI adoption.
Read more:
Now, a new study has found certain uses of AI are driving “cognitive fatigue” and in some cases, a drop in productivity among employees. The study, conducted by Boston Consulting Group and outlined in the Harvard Business Review recently, studied 1488 full-time US workers, delving into their patterns and quantity of AI use.
“We found … cognitive exhaustion from intensive oversight of AI agents ‒ is both real and significant,” said the BCG researchers. “We call it ‘AI brain fry’,which we define as mental fatigue from excessive use or oversight of AI tools beyond one’s cognitive capacity.”
Using AI on genuinely menial tasks to save time actually reduced burnout. But “AI brain fry” ‒ which manifested as mental fog, a difficulty focusing, slower decision-making, and headaches ‒ was particularly high when workers had lots of oversight or monitoring of AI to do. And multitasking using too many AI applications caused mental overload and a drop in productivity.
One finance director described it this way: “I had been back and forth with AI reframing ideas, synthesising data, forming and organising the flow of pillars and work … I couldn’t even comprehend … if what I had created even made sense … just couldn’t do anything else and had to revisit the next day when I could think.”
The BCG researchers found the business cost of this phenomenon was not insubstantial. “Decision fatigue” jumped by a third, raising the cost of suboptimal decision making; error-making increased, and those who reported “AI brain fry” showed an increased inclination to want to leave the job.
Why does it happen?
The Post asked local brain expert, neurosurgeon and University of Otago Professor Dirk De Ridder, whether “AI Brain Fry” is an actual thing.
He says it is.
There are a few models of the way the brain works that explain why. The first of these is the Bayesian brain model, which posits the brain functions as an “inference engine” that uses probability to make sense of the world: “The brain is one big prediction machine that constantly makes predictions and then updates these predictions based on what it samples from the environment by our senses,” De Ridder said.
Then, according to neuroscientist Karl Friston’s “Free Energy Principle”, the brain tries to optimise its model of the world so that it makes as few prediction errors as possible: “Basically, the better your model is, the less prediction errors you make, the more chances you have for survival and procreation.”
Then, you add in “Kolmogorov compressibility”, which says the model has to be as compressed as possible to save the brain energy.
In short, the brain usually finds the 'sweet spot'; a model that is accurate enough to explain what you're seeing, but simple enough (highly compressed) that it doesn't waste metabolic energy.
AI upends that, De Ridder says.
“Because of AI, we get so much information that we need time to compress that information into something our brain can use as a model of the world. If there's too much information, then your model becomes so complex that it can overfit.” Overfitting happens when the brain creates a model of the world that is too complex, and “memorises” noise (or “uncertainty”) or random things instead of identifying the underlying patterns.
Typical over-fitting can be seen in things like patients with PTSD, De Ridder explains, where everything that is linked to a trauma arouses the brain. It is also the reason for the level of anxiety amongst especially young people who are chronically online.
And its impacts are amplified on a brain that is already stressed.
“If you have too much information by AI that cannot integrate in your existing model of the world, then you run into trouble that you overfit your models, and then your predictions become less and less accurate (or in keeping with the real world). And if you make constant prediction errors, your brain says, ‘clearly my model is not good, so I have to withdraw from interactions with the outside world’, and then you become anxious … which can lead to being depressed, or causing more withdrawal. Then if that doesn’t help, you get so tired and can develop chronic fatigue.”
What can be done
De Ridder says if AI is used sparingly, “and only for what’s really essential, then I think it's very good.
“But if you start using it constantly for everything, then your model becomes more and more detached from reality, and then you start making those prediction errors, and then you run into problems of poor decision making, as well as possibly anxiety and depression.”
The BCG researchers say organisations can minimise the AI brain fry amongst workers. They suggest employing AI thoughtfully, including keeping the use of AI agents to just two at a time; setting expectations around AI and workload so employees don’t do themselves damage trying to magic up endless “productivity gains”, and strategically deploying human attention as a finite resource.
Some of the most valuable human skills today, including discernment, decision making, and strategic planning, require focused attention … cultures, teams, and leaders that prioritise cognitive thriving can expect to see better judgements, fewer errors, and higher retention rates for top talent.“