Last October, I woke up to an early rainstorm which I was enjoying at 5:15am while sipping my morning coffee until I felt a water droplet hit my head.
I looked up to realize that water was draining through my kitchen ceiling lights! Not good.
To make a long story short, the project turned into an 8-month nightmare where squirrels had eaten softball sized holes through my roof. To make matters worse, they came back after the initial six-month project had been completed (two days later to be exact) and sounded like a table saw coming through my roof when I was reading bed time stories to my kids.
Without going into the rest of the details, I began using AI to learn as much as I could about roofing issues and squirrel behavior. It became addictive; I wanted it to fix my problem. After all, it knows everything right?
I tried trimming the trees, peppermint balls, and coyote urine. I tried ultrasonic sound wave machines, cameras, and motion sensor lights. I considered a company that would put a layer of concrete around the entire roof (which I could have bought a second home for). I probably called ten different roofing and pest control agencies. I looked up all the local laws in my county.
I had so many questions. Was there a certain type of underlayment (the waterproof plastic like sheet) that they find less appetizing? Why were they doing this? What did they really want? What was the most humane way to deal with this? How effective is it? What are the odds that they could come back? Was my roofer ripping me off? And on and on and on I went.
Every time I thought I had all the answers, I thought of another prompt.
In the process I learned more than I ever wanted to know about squirrels and roofing.
For example, did you know that squirrel’s teeth never stop growing which is why they kept sharpening their teeth on my lead flashing? Did you know that Spanish tile is particularly susceptible to squirrel infestation? Or that red fox tree squirrels are invasive and gray squirrels are protected? I had no idea that squirrels can live 20 years!
I finally hired a wildlife behavior expert who helped me get some answers and solutions.
This was my first experience taking a deep dive into AI.
Some responses were helpful and some just led to more questions. But my overall experience was interesting. I found the use of AI to be strangely addictive, which should not be surprising.
In the modern workplace, AI is taking root in nearly every industry.
The topic is undoubtedly polarizing, and often includes people who are delusionally optimistic about what it can do or those that won’t touch it because they watched too many Terminator movies about Skynet taking over the world.
But more research is coming out to discuss the benefits and costs associated with AI usage at work, along with helpful guidelines to manage the risks.
What does research indicate about AI?
In February of 2026, Harvard Business Review published an article entitled, “AI Doesn’t Reduce Work—It Intensifies It.”
In the study they looked at 200 employees in a US tech company over eight months tracking activity, observing them, and conducting interviews. Importantly, in this study, the use of AI was optional.
Several key findings:
- Work gets faster, but also bigger: More projects at one time and more responsibilities. They finish small tasks quickly so they add something else.
- Scope creep: People begin dabbling in areas outside their primary scope of competence because they have AI assistance and take on tasks maybe they shouldn’t. “I can help with that, I’ll just ask AI.”
- Workday expands: People begin working longer hours because of the first two items above, and think they can squeeze in just one more prompt, which might lead to more questions. And, they often find AI fun and somewhat addictive.
- Higher expectations: People who work faster end up setting up an expectation that all work will be faster, thus having to work faster and longer.
- Cognitive load and burnout increases: Without good checks in place, peoples brains experience increased decision fatigue, cognitive strain, and potential burnout because they are context switching constantly (rapidly jumping back and forth between types of tasks).
- They take less breaks to rest: The space that used to be used for a walk or stepping away from the screen might be used for more interaction with AI. Recovery time is replaced with micro-productivity.
- Since it feels good at first, people keep doing it. Paradoxically, since people had a good quick hit of positive feelings initially, they may keep using it when its not the best option.
In another interesting article, authors point out how “AI leads to brain fry” (March 2026 HBR).
In this article, the study looked at 1,488 US workers across rank and industry and found that although AI can definitely improve productivity in some ways, it can also easily lead to too much interaction and multitasking which leads to mental exhaustion and potential burnout.
The study defined brain fry as: mental fatigue, cognitive overload, brain fog, difficulty focusing, slower and lower quality decisions, headaches, and feeling wired and tired. This was especially linked to using multiple AI programs at once and constantly checking them.
The study found:
- More use is not better. After 3 tools, productivity degrades.
- Oversight was most draining. Using AI for repetitive automated tasks was highly efficient and nearly always saved time, but having to double check its work was exhausting for people.
- Symptoms resemble cognitive overload, not classic burnout. Meaning that people felt unable to think, but not highly stressed or emotionally taxed. “Like having 12 tabs open in your brain.”
- Brain fry increased risk of errors and quitting. 39% more major errors and 39% more likely to quit.
- Some jobs are more susceptible to brain fry than others. Read the article for details.
The bottom line in this article is that AI is not necessarily bad but should be used strategically. They recommended:
- Set clear expectations on AI use and workload.
- Shift metrics from activity to impact.
- Develop worker skill with AI.
- Protect human focus, attention, and creativity, and use it carefully for organizational projects and decisions that really matter.
And lets not forget, AI can be wrong—really wrong. My colleagues and I have found some ridiculous errors over the past year where it essentially hallucinates or straight up makes stuff up. The problem is how confident it can sound while being simultaneously completely incorrect.
Another recent article in Science magazine (March 2026) discussed how sycophantic and addictive AI can be. Here is a quote from the article:
“The model’s responses were nearly 50% more sycophantic than humans’, even when users engaged in unethical, illegal, or harmful behaviors. Users preferred and trusted sycophantic AI responses, incentivizing AI developers to preserve sycophancy despite the risks.”
Next time you want to consult AI, it might be worth asking a friend who won’t tell you what you want to hear. You can also set prompts in your AI account to help combat this.
I’ve written a lot in the past about things like decision fatigue, multitasking, bias, and device addiction, which can be found here in the link to my productivity series. In essence, all those things still apply when considering AI.
My conclusion: Same humans—different tool.
Without focusing on managing ourselves and our habits, many tools can just help us repeat the same bad habits at an accelerated pace. We still need to sleep well, take breaks, single task with focus, and interact with people in order to be our healthiest and most productive selves over the long haul.
The bottom line is that AI can be a productivity amplifier, but is much more helpful when certain guidelines are in place. If these checks are not used, it can easily amplify pressure, increase expectations for more work, and lead to increased fatigue and burnout.
If you or your company are promoting the use of AI, best to do it thoughtfully rather than blindly supporting the “everyone should be using this more” approach.
The wise leader and organization will have many discussions and constant review of how the landscape of AI is affecting their most valuable resource—real humans.
Have a great weekend!
Parker
Resources
- AI leads to brain fry. HBR 2026
- AI doesn’t reduce work, it intensifies it. HBR 2026
- Sycophantic AI decreases prosocial intentions and promotes dependence. Science March 2026.
