The AI Overtime Trap: Why AI Makes You Busier Instead of Less
In this article, you'll learn:
Berkeley just released a study.
The conclusion is alarming: 83% of employees say their workload actually increased after using AI.
They call this phenomenon Workload Creep — AI doesn’t reduce your work, it just gives you the capacity to do more, and then you actually do more.
When I read that number, my first reaction wasn’t worry.
It was: You’re asking the wrong question.
Workload Increased. So What?
Let me start with my own experience.
After using AI, my workload definitely went up.
I probably spend an extra 1–2 hours a day on AI tools — learning new features, adjusting my workflow, reviewing AI output.
But at the same time, a lot of things I used to do by hand just disappeared: data organization, first drafts, format conversion, cross-platform posting…
So the question isn’t “did the workload increase?” It’s: what kind of work increased?
The Industrial Revolution Did This Too
This reminds me of the Industrial Revolution.
After factory machines arrived, human “workloads” didn’t decrease either — new jobs, new demands, new processes appeared everywhere.
But that era raised the standard of living. It didn’t collapse society.
What’s happening with AI right now looks a lot like that.
When tools improve, the result isn’t “doing less” — it’s “doing different things.”
If the extra work time you have now is spent thinking strategically, writing deeper content, making high-value decisions —
that “busier” is a good sign.
If the extra time is spent fixing AI errors, responding to increased demands from your boss, doing a bunch of things you never needed to do before —
that’s the real problem.
It’s Not About Workload — It’s About Flow State
There’s an important concept here: Flow.
Flow is a state of complete focus, where work feels meaningful and time disappears.
After I started using AI, I became busier — but a lot of my work entered a flow state.
Because AI took away the repetitive, distracting, low-brain tasks, leaving only the things I actually want to do.
That feels completely different from being squeezed by your boss.
One is being forced to be busier. The other is choosing to be busier.
People Around Me Who Fell Into the Trap
I’ve seen people who actually got worse after using AI.
They all had the same thing in common:
No goals. Just tools.
They heard ChatGPT was amazing, so they tried dumping work into it. The AI spat out a bunch of stuff, and they still had to spend time editing, verifying, and explaining it to others.
At the end they thought: screw it, doing it myself is faster.
Then they said: “AI is useless.”
It’s not that AI is useless. It’s that they used it with no direction.
My Epiphany: Top-Down vs. Bottom-Up
When I first started using AI, I thought it was “just delegate, AI handles everything.”
After trying it for a while, I realized I had the direction completely backwards.
It’s not about throwing big tasks at AI and expecting complete results.
It’s about starting from the smallest core — planning, delegating, running — then packaging what works into a SOP and letting AI repeat it.
Here’s an example:
My blog workflow right now is Notion draft → Hugo article → English translation → SEO optimization → social post → auto-publish, with AI running at every step.
But this system didn’t get built in a day.
It started with “let AI help me convert the format of one article,” then slowly stacked up until the whole pipeline took shape.
It’s the same concept I talked about in my post about AI Skill testing:
Start small, confirm it works, then scale.
AI Is a Tool, Not a Boss
The root cause of Workload Creep isn’t that AI got stronger — it’s that people and organizations haven’t changed how they use it.
AI delivers more output, so the boss thinks they can ask for more. Employees don’t say no, so the workload actually grows.
That’s not an AI problem. It’s a circle of influence problem — do you have the ability to manage what you actually control?
If all you can control is “how you use AI,” then use it well.
If the problem is your boss or your company culture, AI wasn’t going to save you anyway.
My Advice for Anyone Starting with AI
Not theory — things I’ve actually tried that work:
1. Start With the Task You Hate Most
Don’t start with “the most important task.” Find the thing you hate doing most, the one that wastes the most time and has the least meaning.
Hand that task to AI.
That’s how you actually feel the relief — and get the motivation to go deeper.
2. Build SOPs Starting From the Smallest Task
Don’t try to solve everything at once.
Lock down the process for that one small thing AI helps you with, write it into a SOP, and let it repeat every time.
That SOP is your asset.
3. Track the Time You Save
Once a week, write down: “What did AI help me save time on this week?”
Consciously seeing the saved time is what stops you (or others) from quietly filling it back up.
This is the same idea as the lazy refactor method — understand the structure first, then optimize.
4. Manage Expectations: AI Is Not Magic
AI is powerful, but it’s not all-knowing.
It needs you to give clear direction, needs you to verify outputs, needs you to correct it when it goes off track.
Treat it like a smart assistant, not a machine that can replace your judgment.
FAQ
Further Reading
Lazy Da’s Takeaway
Berkeley says 83% of people are busier after AI. That number is real.
But “busier” doesn’t mean “worse off.”
The question was never how much the workload increased — it’s whether that extra work is actually moving you forward.
The Industrial Revolution made people busier too — but that busyness built the modern world.
AI’s version of busy works the same way.
What you can do:
- Start with the smallest task — don’t try to solve everything at once
- Build SOPs — let AI repeat what works
- Track the time you save — don’t let it quietly disappear
- Tell the difference between “forced busy” and “chosen busy” — the first is a problem, the second is growth
The tool is neutral. The direction is yours.
🚀 已有 1,000+ 讀者加入理財成長之路
