The pitch was clean: artificial intelligence arrives, humans stop doing drudgery, everyone goes home early to touch grass. Tech leaders have spent the last two years selling this vision with the confidence of someone who has never actually used their own product.
Then their employees started talking.
Reports from inside major tech firms paint a different picture — one where AI has somehow created more work, not less. Engineers are now responsible for training models, monitoring outputs, retraining when those outputs go sideways, and explaining to executives why the AI they built cannot actually replace the people building it. The promised efficiency gains have materialized as additional job duties stacked on top of the old ones.
The result: workers clocking 90-hour weeks while their leadership team publishes blog posts about the four-day work week revolution.
This is not a conspiracy. It is something simpler and more human: executives believed their own marketing before the product was ready. They promised shareholders and the market that AI would be a labor multiplier. The technology actually arrived as a labor intensifier — at least for now, at least for the people who have to make it work. Someone has to build the thing, monitor it, fix it when it breaks, and apologize when it hallucinates.
The gap between what tech leaders say AI will do and what their own employees experience doing it is not a bug in the narrative. It is the feature. One group measures success in headlines and stock price. The other measures it in whether they can make dinner plans that do not get cancelled at 6 p.m.
Until AI actually reduces the work, the only thing being disrupted is sleep.