The New York Times last week told the story of Sidharth Hariharan, a mathematics graduate student at Pittsburgh’s Carnegie Mellon University who spent more than two years helping translate one of the past decade’s most celebrated proofs into a form a computer could check, a painstaking task called formalization.

Earlier this year an artificial intelligence system called Gauss, built by the startup Math Inc., finished the job in five days. Hariharan rushed into his adviser’s office in tears. He and his collaborators had been “Gaussed.”

This isn’t a story about mathematicians (they’ll be fine, Jane Street is still out there). It’s about everyone who fears that AI might be able to do their job better than they can, and everyone who leads them.

Hariharan has lost his motivation, and that feeling is coming for everyone who does knowledge work. Leaders have the dual task of reshaping the identities of their people to account for the new way of working while preserving their old skills for the moments when the machines fail.

The person who first identified AI’s threat to morale was a 19th century economist named Karl Marx. You might have heard of him (yes, I’m quoting Marx favorably in capitalism’s in-house organ). Once you strip away all the nonsense (it’s a big project), Marx did have one important insight, which he called alienation.

It’s the idea that industrialization separated workers from the products of their labor and that this changes work from a fulfilling expression of their agency and creativity into meaningless labor. People who argue that AI will not have a devastating effect on the labor market usually note, correctly, that while AI can do parts of many jobs, it so far lacks the capability to do all of almost any job. Even if that remains true, having big parts of your job automated away can be deeply alienating.

But it doesn’t have to be that way. Adding AI to jobs can hollow them out and leave people to babysit a machine that does the interesting part. Or, it can empower people by eliminating drudgery, freeing workers to focus on the uniquely human parts of their jobs that only they can do and improving overall performance. Which way things will go isn’t about technology; it will be decided by who owns the machines and how organizational leaders decide to shape the identities of the people doing the work.

We’ve had a dry run for this transformation in aviation. You’re in more danger when you drive to the airport than when you’re on a plane, and a big part of the reason is that much of the work of flying that plane is now done by computers. There are still pilots up front, but the combination of human and machine is just better at a pilot’s most important job, completing a safe flight, than either could be alone.

That doesn’t mean pilots didn’t feel alienated when the computers started to supplant some of their tasks. In a 1983 paper, “The Ironies of Automation,” the cognitive psychologist Lisanne Bainbridge described how automating the routine parts of a job can leave humans with the hardest parts and the tedium of watching a screen, while the skills they will need in a crisis waste away from disuse. The better the machine, the less the person practices, and the less ready they are when the machine fails.

That’s not a theoretical concern. In 2009, an Air France flight from Rio de Janeiro to Paris flew into a storm, its airspeed sensors iced over, and the autopilot dropped control into the laps of two pilots who had almost never flown a jet by hand at altitude. The protections that normally make an Airbus impossible to stall had dropped out with the failure, and the startled crew stalled it anyway. It fell for three and a half minutes into the Atlantic, recoverable the whole way down. All 228 aboard perished.

Chesley “Sully” Sullenberger, who earlier that same year had lost both engines over New York and glided an airliner onto the Hudson with its protections intact and the reflexes of a former fighter and glider pilot, later asked the industry to consider whether advanced automation makes it harder for pilots to “very quickly intervene and very effectively act when things go awry,” a skill the Air France crew, it turned out, had been allowed to lose. It’s not an accident that one of the bibles of aviation is Wolfgang Langewiesche’s Stick and Rudder. For most pilots right up until the present day, controlling a plane with your hands is what it means to be a pilot. When a computer takes that away most of the time, it’s all too easy to let your skills erode.

So this is the first half of the job facing leaders in the time of AI. Being a pilot had to go from “someone who was gifted at handling stick and rudder” to “someone who got her passengers safely to their destination using every tool available.” Nobody trades in an identity on their own. It took the airlines years of deliberate retraining to make the new one stick. Automation can’t alienate you from your job if that’s how you see your job.

But that’s not the whole problem. Because these tools aren’t omniscient — at least not yet. If you’re just a manager of computer systems, you’ll forget how to use the stick and rudder, or whatever the equivalent of that is in your job. And since these systems tend to be great during normal periods and fail during extreme ones, those skills aren’t going to be useful except at the moments when they’re critical. There are no Sullenbergers among pilots who’ve always relied on their computers.

The aviation playbook comes with a warning label. Pilots kept their seats partly because the law requires someone to sit in them. There is no FAA for knowledge work. If your people keep the skills the machines made optional, it will be because you decided they should.

Hariharan is already working on both parts of the job. He spent months after the scoop cleaning up the machine’s proof, turning its cryptic output into something a person could read, and he says he pulled a real if modest amount of understanding out of the work. His next project will not be a formalization. The next theorem he formalizes, he told the Times, he wants to have proved himself first. Hariharan has found his stick and rudder. You have to help your people find theirs.

____

This column reflects the personal views of the author and does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.

Gautam Mukunda writes about corporate management and innovation. He teaches leadership at the Yale School of Management and is the author of "Indispensable: When Leaders Really Matter."

©2026 Bloomberg L.P. Visit bloomberg.com/opinion. Distributed by Tribune Content Agency, LLC.