We were promised a vacation. The narrative around AI has always been about automation. The idea was that these tools would handle the drudgery—the boring data entry, the repetitive emails—and leave us with more time for high-value creative work. Or better yet, just more time off.
I wanted to believe that. But lately, I’ve noticed that despite using ChatGPT, Gemini, and Claude daily, my day isn’t getting shorter. It’s getting denser.
I dug into a recent article from the Harvard Business Review to understand why. They conducted an 8-month study of a tech company adopting these tools. The results were jarring but validated what I was feeling. They found a 0% reduction in workload.
Here is what is actually happening.
The Data: Pace, Scope, and Intensity
The study found that AI didn’t liberate workers. It intensified their jobs.
When I look at the data, it shows that employees didn’t use the time saved to rest. They used it to work at a faster pace and expand the scope of their responsibilities. This wasn’t necessarily a management mandate. It was self-inflicted. Because the tools make doing more feel possible, we fill every gap with output.
The research identifies three specific “traps” that lead to this intensity.
1. Task Expansion (The Super-Generalist)
This is the most common trap I see. AI lowers the barrier to entry for tasks that used to require a specialist.
In the past, if I needed a Python script to analyze data, I might have asked an engineer or just skipped it. Now, I can ask an LLM to write the code for me. The study calls this the “Super-Generalist Trap.” Product managers start coding. Researchers start doing engineering work.
We see this with “vibe coding.” You can build a tool yourself now. But that means you are now responsible for maintaining it, debugging it, and running it. You go from being a stakeholder to being the builder. You are doing more work that you simply wouldn’t have done before.
2. Blurred Boundaries (Ambient Work)
Prompting is low friction. That is a feature, but it is also a bug for our work-life balance.
Because it takes 15 seconds to ask a question, we do it constantly. I’ve caught myself doing this. I’ll be eating dinner or watching TV, have a random thought about a project, and open ChatGPT on my phone.
It doesn’t feel like “work” because I’m not opening a laptop and logging into a VPN. It feels like texting. But this creates “ambient work.” We eliminate natural pauses in our day because we fill breaks and lunch hours with “quick prompts.”
3. Multitasking (The Partner Illusion)
We treat AI agents like interns, but they still require supervision.
The study highlighted the “Partner Illusion.” We spin up multiple threads. I might have one agent summarizing a PDF, another writing code, and a third drafting an email. It feels productive.
But the cognitive load is massive. You are constantly context-switching. You have to review the output of every single thread because, as we discussed in the “Word Cloud vs. World Model” lesson, these models can hallucinate. You aren’t doing less work. You are managing three different interns who all need your attention simultaneously.
Breaking the Cycle
The cycle is self-reinforcing. Faster work leads to higher expectations. If you deliver a project in 10 minutes that used to take two hours, that becomes the new standard.
So how do we fix it? The answer isn’t to stop using the tools. It is to build an “AI Practice.”
- Intentional Pauses: We need to reclaim the coffee break. Not the “check your phone by the sink” break, but actual disconnection.
- Sequencing: Stop trying to run five parallel processes. Batch your work. Don’t react to every AI output the second it generates.
- Human Grounding: Get back to talking to people. The isolation of talking to chatbots all day distorts your perspective.
Key Takeaways
- Zero Reduction: Data shows AI adoption in tech firms resulted in a 0% reduction in workload.
- Scope Creep: Workers use AI to take on tasks outside their job description, increasing their own burden.
- No Breaks: Low-friction prompting leads to “ambient work” during personal time.
- Cognitive Load: Managing multiple AI threads increases context switching and mental fatigue.
- Solution: You must intentionally schedule pauses and human interaction to break the intensity loop.
Want to earn CPE for this topic?
- Compare Options: See how we stack up against others in our 2025 Flexible CPE Guide
- Understand the Format: Read how Nano-Learning works for CPAs.
- Check Your State: Ensure you are compliant with our State Requirements Guide.
- What is EverydayCPE?
Related Courses:
