2.5 minute read
So, the story goes AI was going to handle all the things that humans didn’t want to do so we could focus on the fun and important stuff, right? So, now that we recall the marketing statement, what is really going on?
So, there’s a study from a workforce analytics firm called ActivTrak. They observed 164,000 workers and noticed a few interesting, noteworthy trends about AI tools being used at work. The following items showed a jump: time spent on email up 104%, time messaging up 145%. Time spent doing focused, uninterrupted work (the kind where you actually solve hard problems) dropped by 9%. So, the engaging stuff we said we would do dropped while all the other time-suck busy work increased?
On our way, in what direction?
Here’s the skinny on another study, this one by researchers at UC Berkeley. They looked inside a tech company with 200 employees. A company where they voluntarily started using AI. It wasn’t a mandate, just if they felt like it. Some did because everything they were reading and hearing said it was helpful. First observations showed it did help.
Workers were getting more stuff done. People started doing things they didn’t before. Non-developers were writing code, and researchers were doing engineering tasks. People felt empowered to do things they were incapable of doing before. Breaks were being replaced by pet projects driven by AI. Less breaks, more work done. Cool right? Everyone’s productivity spikes, and the world is a better place… Or so we thought.
UC Berkeley and Yale researchers have a name for it: workload creep. The time AI saves you does not go back into your pocket. It just gets refilled with other stuff or more work, often by you, without anyone asking you too.
AI servers gets more power, human you don’t
There’s a name for this now. It’s called “AI brain fry” and that’s quite a visual I’m conjuring in my head after hearing about it. Employees who decided the “more the merrier” and chose to use four or more AI tools at the same time started having issues. Decision fatigue set in, more errors were happening. As many will says speed without accuracy, well, it sucks. They don’t really say that, but we sure do think it. These AI agents and chatbots are crazy fast, but humans, well, we have a speed limit and these pesky biological barriers to deal with. I remember reading a book about this called “Your Brain at Work.” Excellent book that said we as humans have a limit in our frontal cortex. The author called them “actors on a stage.” The max we can handle on a state with a bright light shining on them is four, but every time you add one, the accuracy drops. When you are handling the max of four, your accuracy drops like a rock in the ocean.
Workers in the study were hitting an odd kind of exhaustion they hadn’t felt before AI tools entered the picture. As long as you have AI credits or haven’t exceeded the limit on that FREE account, the tools are ready to go thanks to a TON of data centers. You, however, might not be. Push it too hard and you could be staring at the dreaded “burnout.”
Although, I haven’t hit that wall or hopefully never do, there are a few times I looked up from a screen and realized I had been working with that AI chatbot way longer than I should have, all because the next task was right there and the AI made it feel too easy to just keep going. What happened to my natural stopping point? What about hints that it was time to rest? Nope. Beware the endless queue of “one more thing.”
So now what do we do?
Okay, so it’s not all gloom and doom. If we dig deeper the studies did find a “sweet spot” for using AI at work. They called it 7 to 10% of your work hours. Do the math on a 9-hour day that’s about 54 minutes or just short of an hour. Sadly in their study on 3% of AI users were there meaning they either barely used it or OVERDID it.
AI is a good tool, but you are the boss when it comes to working with it. It has no idea what your biological limits are. So, be kind to your human brain. Set a timer, pomodoro if you like. Whatever helps you take some breaks. In-person meetings with other people? Good. A walk here and there is healthy in more ways than one. Staring out a window, no problem, it’s what “people” need to do.
Sources: UC Berkeley Haas School of Business / Harvard Business Review, ActivTrak / Wall Street Journal, Boston Consulting Group, CBS News, Fortune
Tags: AI, Work, Productivity, Burnout
Want to read more AI topics? see below
- Project Glasswing: Anthropic’s Mythos AI Has Hands (And We’re All in Trouble)
- Your Agentic AI Assistant is Coming: 5 Realities to Consider
- Stop Overpaying: The New OpenClaw Strategy
Subscribe to our newsletter!
Get a link to a Free Download: PC Building Mistakes Checklist.
Get my 9-point checklist that prevents costly errors before you buy a single component. Plus other regular tech tips in your inbox...


