A few months ago, in a board meeting, I heard a sentence that kept spinning in my head: "With AI, we'll finally be able to do the work of ten people with three." The tone was celebratory. The room agreed. I agreed too β at the time. It was only later, in the car, driving home at nine in the evening, that I realized what I had agreed to: that the goal of the most transformative technology of our generation was, essentially, to fire seven people and overload three.
I'm not being naive. I know spreadsheets, quarterly targets, and the pressure to deliver results for shareholders who live on the other side of the world. I've been working with this for over twenty years. But it's precisely because I know it from the inside that the sentence bothered me. Because I've already seen what happens when "doing more with less" stops being an aspiration and becomes culture: people get sick. Silently, almost always. Spectacularly, sometimes.
Byung-Chul Han calls this the burnout society β a system where violence doesn't come from outside, but from within, from the person who demands too much of themselves. The individual self-exploits. They answer emails at eleven at night not because the boss told them to, but because they "need to" β and they no longer know who defined that need.
Now place artificial intelligence in this scenario.
If AI is used only to accelerate what already exists β more tasks, more deliverables, more "productivity" β it won't liberate anyone. It will be yet another layer of self-exploitation. The professional who used to take four hours to produce a report now takes forty minutes. Wonderful. But what happens with the three hours and twenty minutes that are "left over"? Does anyone believe they'll meditate, have a calm lunch, or pick up their kids from school? No. They'll produce four more reports. And feel guilty when they can't manage five.
Jung would call this inflation of the persona β that mask we wear to function in the professional world and which, when it works too well, begins to swallow everything behind it. The person becomes the job title. The executive forgets he's a father. The manager forgets she likes to paint. The analyst forgets they were curious about something that wouldn't fit in an OKR. And AI, in this context, becomes a persona amplifier: it makes you seem more productive, more complete, more available β without anyone asking at what cost.
In my doctoral research at FGV, I investigate digital disconnection policies β the right of people not to be contacted outside working hours. It sounds simple. It is anything but simple. Because disconnecting isn't just turning off your phone. It's confronting a culture that confuses presence with commitment, speed with competence, and permanent availability with loyalty.
I talk to HR managers who know this is a problem. Many have the data: sick leave for burnout rising, high turnover, climate surveys with troubling numbers. But the solution they find is almost always cosmetic: a meditation app in the benefits package, a mindfulness talk at the safety week, a nice post for mental health awareness month. Treating the symptom without touching the cause is comfortable. Changing the culture is another thing entirely.
That's where AI could do something different.
Imagine using artificial intelligence not to produce more, but to protect time. Tools that automatically block communications outside working hours. Systems that identify overload patterns before they become sick leave. Analyses that show, with data, that the team that rests performs better than the one doing chronic overtime. None of this is science fiction. The technology exists. What's missing is intention.
And intention is a question of leadership, not IT.
I study and practice AI every day. I use it to automate what's mechanical, to think better about what's complex, to gain time that I reinvest in things that matter. But I do this with a conviction that isn't technical β it's philosophical: technology that doesn't improve people's lives is just sophisticated noise. And "improving life" isn't synonymous with "increasing output." It is, many times, the opposite: it's allowing people to stop.
I know this is an uncomfortable position for those who live by quarterly results. But the bill comes due. It comes in the sick leaves, in the turnover, in the loss of good people who simply got tired. And when it comes, it's always more expensive than prevention would have been.
What to do with the algorithm's optimizations? I don't know. But I didn't want to let it decide. Whoever decides should be alive, should enjoy life β not optimize it.