In November 2024, I wrote “The Great AI Lie: Greater efficiency does not always equate to improved mental health.” I argued that AI, like most major technological advancements, would not give us our time back. Instead, efficiency gains would be quietly reinvested into more work, more output, more grind. I pointed out how previous technological shifts like the internet and cell phones had resulted in a net loss in work/life balance. This was not to discount the value they have added to our lives, just to frame the cost.
I wanted to be wrong. Nothing would have brought me more joy than to write this follow-up to mark the celebration of a fire service that had taken those reclaimed hours and spent them on their families, their health and their crews. That is not the article you are reading, though.
| READ NEXT: AI can support the fire service — but only if safety leads
What the research now shows
Over the past 18 months, researchers have put numbers to what many of us were feeling. An eight-month workplace study out of the University of California – Berkeley, published in Harvard Business Review, reached a conclusion that can be understood by the title alone, “AI doesn’t reduce work — it intensifies it.” Employees using AI worked at a faster pace, took on a broader scope of tasks, and extended work deeper into their evenings. Nobody told them to; they did it instinctively.
If that sounds familiar, it should. It’s the same instinct I watched in that National Fire Academy classroom when I asked a room full of new executive fire officers what they’d do with 10 extra hours a week. After a full week of conversations about the health of the fire chief, nearly every answer was more work.
Other findings have stacked up alongside it. Researchers coined the term “AI brain fry” to describe the cognitive exhaustion that comes from excessive use and oversight of AI tools beyond our mental capacity. Workers experiencing it reported more mistakes, slower decisions and deeper fatigue. Surveys of AI-enabled workers show dramatically higher burnout rates, not lower ones. And “workslop” entered the workplace vocabulary: AI-generated content that looks like work, reads like work, but quietly shifts the real effort onto other coworkers who now feel pressure to do something with it.
The insight illusion
AI isn’t always surfacing hidden problems. Most of the time, the problems weren’t hidden at all. They were sitting in plain view; we had simply judged them as not important enough to act on. There’s a difference between undiscovered and deprioritized, and AI has a way of blurring it.
When an AI tool flags 40 “insights” from your incident data, your training records or your inspection backlog, it isn’t telling you what matters. It’s telling you what’s there. The judgment about what matters — you know, that thing we actually get paid to do as fire service leaders — still belongs to us. But every flagged item now demands a decision, an explanation or, at minimum, another dismissal.
Summaries are dissolving our ability to think critically
Then there’s the quieter cost, the one I worry about most. AI summaries of everything — whether it’s articles, incident reports, policy updates, meeting notes — are capturing the little time we had left for true, in-depth reading. That reading and thinking is where the value of our experience lives, where years on the job tie uniquely to the details of a real problem. Now an AI model without 20 years of fire service experience provides you with a summary of those details, but just the ones it thinks are important.
Think about what deep reading actually is. It’s not information transfer; it’s consideration. It’s the pause on page four where you connect what you’re reading to that call last month, to a conversation with your crew, to a problem you’ve been chewing on for a year. The summary delivers the conclusion but skips the consideration. We are consuming more content than ever and thinking about less of it.
In a profession where judgment is the product, where the quality of a decision on the worst day of someone’s life is the whole job, is outsourcing the thinking part of our reading really an efficiency gain?
The pressure trap
There’s one more force at work, and it’s the one nobody wants to say out loud: fear. Across the fire service, people are sensing that failure to embrace AI will leave them behind. The uncomfortable part is, they may be right.
But look at the trap we’ve built. We tell people with completely full plates that they must also find time to learn AI. AI changes monthly, by the way, so with what hours should they be keeping up with this? The very technology sold as the solution to our time problem has become a new demand on our time, and the anxiety of keeping up has become its own form of workload. We’ve created a profession-wide unfunded mandate, being paid in nights and weekends — those same nights and weekends we should be protecting for our families and loved ones.
Back to the mountain
In Part 1 of “The Great AI Lie,” I borrowed Dr. Kenneth Kamler’s Everest lesson: The goal of climbing Mt. Everest isn’t reaching the summit, it’s making it back down alive. I extend this as a metaphor for our fire service careers. We focus so much on the climb that we sacrifice the very things we need to make it down the mountain safely. The ironic thing is that AI truly can be a transformative technology that helps us to both rise to the summit and return safely, but that requires a specific intention — a deliberate, written-down, talked-about decision regarding how your organization will and won’t use these tools, and how you will and won’t allow it to impact the lives of your people.
Most departments haven’t had these conversations. Industry surveys show AI use is widespread while AI governance, the actual policy, leadership discussion, and training plans lag far behind. Perhaps that’s because AI can’t have those difficult conversations for you.
So what do we do?
I’m not anti-AI. I use it, I teach it, and I have conversations about it at every level of our industry. But using a tool wisely starts with naming what it’s doing to you. Here is where I’d start:
- Put it in writing. If your department uses AI, your department needs an AI policy. Not a standardized template but yours, unique to your organization and your values — what it’s for, what it’s never for and who’s accountable.
- Protect the saved hours. This one is tough, and we’ve lost this fight before. (If you think we haven’t, then tell me you haven’t checked an email from your phone during dinner with your spouse or brought your work laptop home on the weekend to knock just a little bit of work out before Monday. I know that answer for most already.) As a leader, it is up to you to set a culture around AI efficiency. You need real answers on how much efficiency you are gaining from AI and then set clear expectations around how those hours should be reinvested. I can’t tell you what that looks like for your organization, only you can.
- Guard the deep work. Decide what work must be human led, for example, after-action reviews, mayday debriefs, LODD narratives, anything safety-critical or trust-critical. Some things should cost us time — that’s the point of them.
- Make learning AI part of the job. If we expect people to build this skill, give them duty time to build it. You wouldn’t ask your people to take a new extrication tool home over the weekend and learn it, don’t do it with work-focused AI. Bring in classes and find internal AI champions — you likely have resources available to you already.
Walking the talk, again
In the first article, I admitted that AI helped me draft it, and I promised to spend the saved hours with my family instead of taking on more work. I can say that I kept that promise, albeit imperfectly, which is the way we are forced to keep most promises that matter.
AI helped with this one, too. But the ideas in it came from a place no model can reach, meaningful conversations, long reads I refused to summarize, and 18 months of watching this profession I love struggle to find its footing with this technology. The lie hasn’t changed and neither has the choice.
REFERENCES
Ranganathan, A., & Ye, X. M. (2026, February 9). “AI Doesn’t Reduce Work—It Intensifies It.” Harvard Business Review.
Bedard, J., Kropp, M., Hsu, M., et al. (2026, March 5). “When Using AI Leads to ‘Brain Fry.’” Harvard Business Review.
Niederhoffer, K., Rosen Kellerman, G., Lee, A., et al (2025, September 22). “AI-Generated ‘Workslop’ Is Destroying Productivity.” Harvard Business Review.
The Upwork Research Institute. (2024, July 23). “Upwork Study Finds Employee Workloads Rising Despite Increased C-Suite Investment in Artificial Intelligence.” Upwork Inc.
