A new piece in Harvard Business Review makes a bold claim: AI doesn’t reduce work. It intensifies it. Ranganathan and Ye (2026) studied how knowledge workers use generative AI in their daily routines and found that, contrary to the popular narrative, AI isn’t freeing up time. It’s quietly expanding workloads, blurring boundaries, and pushing people into unsustainable patterns.
The findings are provocative. And some of them are genuinely important. But I think the picture is more complicated than the article suggests.
The Case for AI as a Work Intensifier
The core argument from Ranganathan and Ye is that AI lowers the barrier to starting tasks so effectively that people end up doing more, not less. One participant in their study put it bluntly: “You had thought that maybe, oh, because you could be more productive with AI, then you save some time, you can work less. But then really, you don’t work less. You just work the same amount or even more”.
That observation tracks with something many of us have felt. AI makes it easy to say yes to one more task. A quick prompt during lunch. A draft started between meetings. A coding problem tackled late at night because it suddenly feels manageable. Each small action feels harmless on its own, but they accumulate. The natural pauses that once provided recovery, the walk between meetings, the quiet evening, gradually disappear.
The authors also flag what is known as role creep. AI enables people to step into tasks that used to belong to other roles. Designers start writing code. Product managers troubleshoot technical issues. Researchers take on engineering work. This feels empowering in the moment, but it scatters attention and creates hidden labor. Someone still has to review, verify, and correct the AI-assisted output, and that oversight work often falls on colleagues who didn’t ask for it.
Ranganathan and Ye are direct about the trajectory: “without such practices, the natural tendency of AI-assisted work is not contraction but intensification, with implications for burnout, decision quality, and long-term sustainability.”
Does AI Increase Productivity? A Different Perspective
Here’s where I part ways with the article, at least partially.
The argument that AI increases workload assumes a particular relationship with the technology: one where AI generates more tasks and the human tries to keep up. And yes, that pattern is common. But it’s a pattern of use, not an inevitability. The question of whether AI increases productivity depends entirely on how you use it.
I’ll use my own work as an example. I write, research, and publish across multiple platforms every week. When I use AI, I don’t use it to generate more work. I use it to handle the mechanics so I can focus on what actually matters: thinking, reading, forming opinions, and connecting ideas. AI takes care of formatting, language polishing, technical SEO, and structural edits. That frees my mental energy for the creative and analytical work that only I can do.
That’s a fundamentally different dynamic from what Ranganathan and Ye describe. In their study, AI becomes a treadmill. In my workflow, AI becomes a filter that separates deep work from busywork.
I wrote about this exact experience when I described how AI agents handled the entire technical backend of my three websites in a single day. Tasks that used to cost thousands of dollars and weeks of back-and-forth with freelancers were completed in hours. That’s genuine productivity, not intensification. The key difference was intention. I knew what I wanted AI to do, I set the boundaries, and I stopped when the work was done.
The Cognitive Cost Is Where It Gets Serious
Where I fully agree with Ranganathan and Ye is on cognitive load. They describe workers who feel productive but not less busy, who juggle multiple AI outputs simultaneously, and who report increasing mental fatigue despite getting more done. As they write: “What looks like higher productivity in the short run can mask silent workload creep and growing cognitive strain.”
This connects to a growing body of research on what AI does to our thinking. Bai, Liu, and Su (2023) documented how reliance on ChatGPT affects learning and memory, showing that offloading cognitive tasks to AI weakens the mental processes that make us learn. Gerlich (2025) extended this further, arguing that widespread cognitive offloading through AI tools is eroding critical thinking capacity at a societal level. And Kosmyna et al. (2025) found measurable cognitive debt in participants who used AI for writing tasks, meaning the brain literally does less processing when AI handles the intellectual heavy lifting.
I’ve covered this cognitive dimension in a previous post on how AI is quietly reshaping the way we think, where Shaw and Nave (2026) introduced the concept of cognitive surrender. The pattern is consistent across these studies: when AI takes over cognitive work without the user maintaining active engagement, thinking weakens.
But here’s the nuance that matters. Cognitive offloading becomes dangerous when it replaces thinking. When AI handles the mechanics of a task and the human stays fully engaged with the ideas, the outcome is different. The cognitive cost the research describes is a risk, not a certainty. And it’s a risk that can be managed with the right habits and boundaries.
AI and Productivity Need Intentional Boundaries
Ranganathan and Ye close with a recommendation I fully support. They argue that organizations need what they call an “AI practice,” a set of shared norms that guide how AI is used, where work should stop, and how pace is regulated. Their final observation captures it well: “Our findings suggest that without intention, AI makes it easier to do more but harder to stop.”
That line deserves to be printed and taped to every knowledge worker’s desk.
The same principle applies to educators. If we bring AI into classrooms or professional workflows without deliberate structure, we’ll see the same pattern: more output, less reflection, and growing fatigue masked as productivity. The question of whether AI increases productivity has a clear answer, but only when the conditions are right. AI increases productivity when it’s used with intention, directed toward specific goals, and paired with the discipline to stop when the work is done.
Without that intention, it just makes us busy.
References
- Bai, L., Liu, X., & Su, J. (2023). ChatGPT: The cognitive effects on learning and memory. Brain and Behavior, 13(10), e30.
- Gerlich, M. (2025). AI tools in society: Impacts on cognitive offloading and the future of critical thinking. Societies, 15(1), 6.
- Kosmyna, N., Hauptmann, E., Yuan, Y. T., Situ, J., Liao, X. H., Beresnitzky, A. V., & Maes, P. (2025). Your brain on ChatGPT: Accumulation of cognitive debt when using an AI assistant for essay writing task. arXiv preprint arXiv:2506.08872.
- Ranganathan, A., & Ye, X. M. (2026, February 9). AI doesn’t reduce workâit intensifies it. Harvard Business Review. https://hbr.org/2026/02/ai-doesnt-reduce-work-it-intensifies-it