AI Deleted the Pauses | Eliran Keren

Something moved this month, and most of the coverage is reading it wrong.

Across a run of July reports on AI and work, the same finding keeps surfacing: mental fatigue has passed workload as the strongest predictor of burnout. Not how much you have to do. How much you have to hold. The volume of work didn’t spike. The kind of tiredness did.

And then the part that should stop you. The workers reporting the most burnout are the ones getting the most out of AI.

The paradox in the data

Boston Consulting Group ran the numbers for Harvard Business Review across close to 1,500 US workers. The ones with heavy AI-oversight responsibilities reported 14% more mental effort, 12% more mental fatigue, and 19% more information overload than the workers who barely touch it. BCG has a name for the top of that curve: AI brain fry. Mental fog, an inability to focus, decisions that arrive slower. It shows up worst in the people leaning on the tools hardest.

Read that again. A productivity tool whose heaviest users end up more tired, not less.

The prescriptions write themselves. Time-box your AI sessions. Take more breaks. Managers, be intentional about it. Every piece of coverage lands on some version of pace yourself. It’s the same instinct that answers a flooding basement by handing out towels. Not wrong. Aimed at the wrong layer.

What actually changed

AI didn’t pile on more work. It deleted the pauses.

Knowledge work used to come with recovery built in. The task that took twenty minutes — formatting the deck, drafting the summary, pulling the numbers — wasn’t only output. It was also the gap. The low-intensity stretch where your mind caught up with itself before the next hard thing. Psychologists interviewed by Fortune this spring made the point plainly: the “mindless” tasks AI is so good at removing are exactly the ones the brain uses to recover.

Now the twenty-minute task takes twenty seconds. The gap is gone. You go from one high-intensity decision straight into the next with nothing in between, and the nervous system never leaves alert. That’s the fatigue. Not more tasks. No pauses between them.

But there’s a layer underneath even that, and it’s the one nobody is naming.

The pause was doing something

The pause was never just rest. It was reconciliation.

It was the minute where you confirmed the thing you just did was actually done — the number was right, the email went, the hand-off landed — and closed that loop before opening the next one. Quiet, automatic, invisible. You never counted it as work because it never felt like work.

AI removed the minute and kept the loop open. Every hand-off to an agent finishes the task and leaves a question hanging. Did it do what I asked. Is that figure real. Did the downstream thing fire. What did I just agree to by accepting this. The work gets offloaded. The tracking gets uploaded — to you.

So the record lives in the one place it was never supposed to: your head. Every open loop the AI generated, held by the only part of the system with no ledger and no undo. At machine speed. With no pause left to write any of it down.

That is the burnout. Not the volume of decisions. The weight of the open ones, carried by a person because nothing else is carrying them.

Why breaks can’t fix it

This is why the wellness prescription can’t reach the problem. You can’t rest your way out of holding two hundred unresolved commitments. The load isn’t on your calendar. It’s in the context you’re keeping that no system is keeping for you.

Pace the human all you want. The loops don’t close on a break. They wait for you to come back and pick them up again.

Burnout in the AI era is an infrastructure gap wearing a wellness costume. The tools got faster at producing open loops and nothing got built to hold them, so they pooled in the last place with any memory: the person.

The pauses are not coming back. Tasks will keep collapsing from twenty minutes to twenty seconds. That’s the entire point of the tools, and it isn’t reversing. The only open question is where the loops those tasks used to close are supposed to go.

Right now they go into people. That’s what the burnout numbers are measuring. Not weak workers. A missing layer.

Somewhere to put the open loops that isn’t a human’s memory. That isn’t a wellness program. It’s infrastructure.

That’s what we’re building.