Essentially the most seductive narrative in American work tradition proper now isn’t that AI will take your job. It’s that AI will prevent from it.
That’s the model the business has spent the final three years promoting to thousands and thousands of nervous people who find themselves keen to purchase it. Sure, some white-collar jobs will disappear. However for many different roles, the argument goes, AI is a power multiplier. You turn into a extra succesful, extra indispensable lawyer, guide, author, coder, monetary analyst — and so forth. The instruments give you the results you want, you’re employed much less exhausting, everyone wins.
However a new study revealed in Harvard Enterprise Assessment follows that premise to its precise conclusion, and what it finds there isn’t a productiveness revolution. It finds firms are prone to changing into burnout machines.
As a part of what they describe as “in-progress analysis,” UC Berkeley researchers spent eight months inside a 200-person tech firm watching what occurred when employees genuinely embraced AI. What they discovered throughout greater than 40 “in-depth” interviews was that no person was pressured at this firm. No one was instructed to hit new targets. Folks simply began doing extra as a result of the instruments made extra really feel doable. However as a result of they might do this stuff, work started bleeding into lunch breaks and late evenings. The workers’ to-do lists expanded to fill each hour that AI freed up, after which saved going.
As one engineer instructed them, “You had thought that possibly, oh, since you could possibly be extra productive with AI, you then save a while, you possibly can work much less. However then actually, you don’t work much less. You simply work the identical quantity or much more.”
Over on the tech business discussion board Hacker Information, one commenter had the same reaction, writing, “I really feel this. Since my workforce has jumped into an AI the whole lot working type, expectations have tripled, stress has tripled and precise productiveness has solely gone up by possibly 10%. It appears like management is placing immense strain on everybody to show their funding in AI is price it and all of us really feel the strain to attempt to present them it’s whereas truly having to work longer hours to take action.”
It’s fascinating and likewise alarming. The argument about AI and work has all the time stalled on the identical query — are the positive factors actual? However too few have stopped to ask what occurs when they’re.
Techcrunch occasion
Boston, MA
|
June 23, 2026
The researchers’ new findings aren’t fully novel. A separate trial final summer time discovered skilled builders utilizing AI instruments took 19% longer on tasks whereas believing they had been 20% sooner. Across the similar time, a Nationwide Bureau of Financial Analysis research monitoring AI adoption throughout 1000’s of workplaces discovered that productiveness positive factors amounted to just 3% in time savings, with no important affect on earnings or hours labored in any occupation. Each research have gotten picked aside.
This one could also be tougher to dismiss as a result of it doesn’t problem the premise that AI can increase what staff can do on their very own. It confirms it, then exhibits the place all that augmentation truly leads, which is “fatigue, burnout, and a rising sense that work is tougher to step away from, particularly as organizational expectations for pace and responsiveness rise,” in line with the researchers.
The business wager that serving to individuals do extra can be the reply to the whole lot, however it could transform the start of a unique drawback fully. The analysis is price studying, here.
