Grind culture gets an AI upgrade

TL;DR: An ongoing University of California, Berkeley study embedded researchers in a US tech company to look at their AI use. Yesterday, they announced a striking finding: AI didn't lighten workloads—it "intensified" them, making people work faster, take on more tasks, and blur the line between work and everything else. It's a pattern we’ve seen before: When technology makes effort cheaper, we don't do less of it—we end up doing more.

What happened: Maybe you've dreamed of a day when you can tell your AI assistant, "Do all my work for today. Make zero mistakes." It sends emails, attends meetings, crunches all the data, and even schmoozes with the boss. You kick back and relax.

But according to a new study from UC Berkeley, AI isn't reducing people's workloads. It’s piling even more on their plates. Researchers spent eight months last year embedded at a US-based, 200-person tech company, observing how AI changed daily work. Nobody was forced to use it—the company just handed out enterprise subscriptions and let workers figure it out.

Empowered with tools that made “‘doing more’ feel possible,” the workers became the type of employees recruiters demand in job descriptions: go-getter rockstar ninjas wearing way too many hats. Product managers started writing code. Designers tackled engineering tasks. Software engineers spent more time reviewing the AI-assisted work their "vibe coding" colleagues now produced. Breaks, meanwhile, shrank—workers got into the habit of sending one last prompt before lunch or after hours. Employees at this firm became masters of multitasking, running parallel AI tasks the way a short-order cook juggles tickets.

Why it matters: These early results show how AI’s potential for supercharging productivity might be a monkey's paw wish come true, especially for workers. You don't get your time back to pursue projects you're more interested in, or even clock out early—you take on more work that previously "justified additional help or headcount," says Harvard Business Review.

Companies thrilled by the prospect of more work for less pay might want to slow down, too. HBR warned that, in the long run, the initial "productivity surge" can result in sloppier work and even burnout—and as the frustration mounts, workers might quit.

The ways AI might affect your job aren’t a far-off reality. As of now, about 12% of US workers are using AI every day at their job, per Gallup. AI job redistribution is happening, too: Over 50,000 US job cuts were attributed to AI in 2025. That doesn’t mean AI can really take over these jobs—many companies blaming it for layoffs don’t appear to have the tools needed to truly replace workers—but the end result is that a smaller crew of human employees are already taking on more work under the guise that AI can automate tasks and boost productivity.

Productivity-maxxing through history: If we look at past waves of technological change, the pattern that emerged at this tech firm isn’t surprising. A more efficient steam engine didn’t reduce coal consumption—it increased it. The invention of the power loom boosted weaving productivity many times over—textile workers didn’t labor less. They worked even harder in factories that could now demand much higher production quotas. And new tech doesn’t just usher in more work, but new tasks. Self-checkout at grocery stores didn’t eliminate human cashiers, it just meant workers would oversee these finicky machines on top of their other duties.

The hamster wheel never stops: The UC Berkeley researchers call for companies to develop an "AI practice"—norms around when and how to use AI that might stem some of this workload creep. But don’t hold your breath expecting companies to voluntarily slow down a productivity gold rush. —WK