You bought the tool that was going to give you your evenings back. For a week or two it did. Then the saved time filled in, like a footprint in wet sand, and you were as busy as before, only now with better tools.
You are not imagining it, and you are not weak. You are living inside a pattern almost a hundred years old.
What happens to the time AI saves?
It goes back into the work. When a global jobs platform asked more than eighty thousand workers across eight countries what they did with the time AI saved them, the most common answer was not “left earlier.” It was “took on other tasks and projects.”1 Enterprise workers with AI tools report saving forty to sixty minutes a day.2 The minutes are real. The default destination is more work.
A careful study makes the mechanism vivid. Researchers spent eight months watching a two-hundred-person company adopt AI. Nobody was forced to use it. People embraced it because doing more felt easy and rewarding, and then, quietly, the work expanded. They moved faster, took on a wider range of tasks, and stretched into more hours. No one asked them to. The workloads crept upward until they became overwhelming.3
Why does every new tool do this?
Because saved time is just capacity by another name, and in a culture that treats idle time as waste, capacity gets filled. The electric motor, the computer, and email all promised to hand us time. Each was converted, with remarkable efficiency, into more work. The ratchet only turns one way.
AI is the largest efficiency windfall in a hundred years, arriving faster than any predecessor, and it is landing on a working culture engineered to waste it. The defaults are not neutral. They all point at more.
Could AI actually shorten the week?
Yes, but only on purpose. The four-day-week trials prove the conversion is possible: cut the hours, keep the pay, and output holds while burnout falls.4 The lesson is that shorter hours come from re-engineering the structure of work, not from the tool itself.
So the question that matters is not whether AI makes you faster. It will. The question is whether you let the windfall make you busier, like all the others, or whether you become one of the people who finally takes it as time.
What does taking it as time look like?
It is mechanical, not heroic. Decide what “enough” looks like for a piece of work before you start. Produce that, and stop. Bank the saved time as time, deliberately, before the work expands to swallow it. Cap your own output so a faster task does not silently become a longer report. Put the reclaimed block on the calendar and defend it like an appointment.
None of this happens automatically. Left alone, the forces that eat saved time are stronger than good intentions. But they are a list of problems with solutions, and the solutions are the subject of this book. AI will go the way of every other efficiency wave by default. The only thing that changes the default is you.
Frequently asked questions
Will AI reduce our working hours?
Not by default. Surveys show the most common use of AI-saved time is taking on more tasks, not leaving earlier. Every previous efficiency wave was converted into more output rather than leisure, and AI is following the same pattern unless people deliberately bank the saving as time.
How much time does AI actually save?
Enterprise workers report saving roughly 40 to 60 minutes a day, and early assistant users self-report around 1.2 hours a week. The minutes are real. The question is where they go, and the default is back into work.
Can AI give us a shorter week?
It can, but only if the saving is converted into time on purpose. AI is the largest efficiency windfall in a century, but the defaults all point at more output, so the outcome depends on individual and collective choices, not the technology itself.
About the author
Tom Goodwin is the author of Don’t Work Harder, a book about taking the time AI gives back as time rather than more work. He is a co-founder of GAMEPLAN and writes on productivity, technology, and the economics of the working week.
Footnotes
Footnotes
-
Indeed Hiring Lab survey of more than 80,000 workers across eight countries, December 2025. ↩
-
Goldman Sachs AI Adoption Tracker, drawing on OpenAI enterprise data, December 2025: average savings of 40 to 60 minutes per day among AI users. ↩
-
Aruna Ranganathan and Xingqi (Maggie) Ye, “AI Doesn’t Reduce Work, It Intensifies It,” Harvard Business Review, February 2026, reporting an eight-month field study at a roughly two-hundred-person technology firm. ↩
-
Autonomy and 4 Day Week Global, The Results Are In: The UK’s Four-Day Week Pilot (February 2023): output broadly held while burnout fell, with 92 percent of companies continuing the four-day week. ↩