Alex Prompter@alex_prompter
🚨 BREAKING: Berkeley just proved that AI doesn’t save you time. It makes you work MORE.
Researchers Aruna Ranganathan and Xingqi Maggie Ye from Berkeley’s Haas School of Business spent 8 months embedded inside a 200-person tech company. Twice-weekly observations. 40+ deep interviews across engineering, product, design, and operations.
This wasn’t a survey. They watched what actually happens when a company gives everyone AI tools and says “go.” What they found contradicts everything AI vendors have been selling you. Employees worked at a faster pace, took on a broader scope of tasks, and extended work into more hours of the day. Nobody asked them to. The company didn’t even mandate AI use. People just voluntarily did more because AI made “doing more” feel possible.
One employee put it perfectly: “You had thought that maybe you save some time, you can work less. But then really, you don’t work less.”
That quote should be taped to every monitor running Cursor, Claude, or ChatGPT right now.
And a 2024 Upwork study backs it up: 77% of employees using AI said the tools had actually INCREASED their workload. Nearly half didn’t even know how to achieve the productivity gains their employers expected.
The researchers found 3 patterns destroying work-life balance. First, task expansion. Product managers started writing code. Researchers took on engineering work.
The scope of “my job” widened because AI made everything feel doable. Hiring got postponed because employees absorbed work that would have justified new headcount. Second, blurred boundaries.
Workers sent prompts during lunch, before meetings, at 9pm. AI dropped the friction of starting any task to near zero, and natural stopping points in the workday just dissolved. Third, cognitive overload.
People ran multiple AI agents simultaneously while reviewing code, drafting docs, and sitting in meetings. Both human and machine constantly in motion.
Here’s the cycle that traps you. AI speeds up a task → expectations for speed rise → you rely more on AI → you take on wider scope → workload intensifies → repeat.
The researchers call it “workload creep.” No manager told anyone to work harder. The tools just made doing more feel accessible and rewarding. So people kept going until they couldn’t.
The most dangerous part: in the moment, it feels amazing.
Workers described momentum, expanded capability, the thrill of building things they never could before. But when they stepped back and looked at the full picture, they felt busier, more stretched, unable to disconnect.
By month 6 of the study, reports of burnout, anxiety, and decision paralysis had spiked. Short-term momentum. Long-term strain.
There’s also a competitive dynamic nobody talks about. When your colleague uses AI to take on extra responsibilities, standing still feels like falling behind. Nobody formally raises expectations. But informal norms shift fast. Within months, doing what AI makes possible becomes what’s expected.
The people who set healthy boundaries start looking like underperformers. That’s a toxic dynamic where sustainable work becomes career-limiting.
The researchers propose something they call “AI Practice.” Not “use AI more” or “use AI less.” Intentional habits. Structured reflection intervals built into workflows, not “take breaks when you need to” because nobody does.
Scheduled reviews where teams assess if AI-enabled expansion has crossed sustainable limits. Clear guidelines on when NOT to use AI and which tasks shouldn’t expand just because they can.
I felt this in my own workflow. AI gives you superpowers. But superpowers without discipline just mean you never stop working. The fix isn’t to stop using AI. It’s to stop letting AI decide how much work you do.
Set the scope BEFORE you prompt.
Define “done” BEFORE the tool makes everything feel possible.
