AI Is Making Work More Intense, Not Less. Research Shows Why That Gap Will Widen.

The Shorter Workweek Never Showed Up

The 15-hour workweek prediction was wrong. Again.

Researchers at ActivTrak analyzed the digital activity of more than 10,000 workers and found that AI adoption made work life more intense, not easier. Time spent on email, messaging, and chat apps more than doubled among early AI adopters. Use of business software rose 94 percent.

Researchers at UC Berkeley's Haas School of Business found a parallel pattern: workers using AI started reclaiming tasks they had previously outsourced, because coding, engineering, and similar work became easier to do themselves. They squeezed in work bursts on evenings, weekends, and idle moments whenever AI was accessible. Multitasking surged as workers supervised multiple bots simultaneously.

Every labor-saving technology in modern history has prompted people to do more, not rest. Planes didn't reduce how much people traveled. Cars didn't shrink commutes. They expanded where people were willing to live.

The Real Dividing Line

The ActivTrak data carries one striking number: focused, uninterrupted work fell by 9 percent among AI adopters. Researchers have started calling the resulting mental state "AI brain fry." More output, more fragmentation, less depth.

The technology is creating two forces pulling in opposite directions at the same time. People are accomplishing more volume while concentrating less. That tension will matter a lot depending on what kind of work you do.

Psychologists have a term for people who genuinely enjoy hard thinking: high need for cognition. These are the people who seek out difficult problems, read dense material for pleasure, and treat intellectual struggle as inherently rewarding. On the other end of the spectrum are what researchers call cognitive misers — people who find effortful thinking unpleasant and avoid it when given the option.

AI, by making cognitive shortcuts cheaper and easier, is essentially handing cognitive misers a permanent off-ramp. That's a problem.

When Intelligence Is Cheap, Volition Becomes the Scarce Resource

The strongest case for AI optimism is real: barriers to entry are collapsing. A solo developer can now build software that would have required a team. A small business owner can produce marketing copy, legal summaries, and financial models without hiring specialists. That democratization is genuine and worth acknowledging.

But the concern that deserves equal weight is this: if AI does the cognitive heavy lifting by default, and if workers let it, the people who relied on wrestling through hard problems to sharpen their thinking will stop getting that workout. Skills atrophy. Judgment, which develops through repeated trial and error under pressure, doesn't accumulate if the AI is always making the first cut.

The UC Berkeley Haas findings suggest that workers who thrived with AI weren't the ones who handed tasks off entirely. They were the ones who used AI to expand what they could attempt and then stayed engaged in the output. The technology amplified their existing drive. It didn't manufacture drive where none existed.

Knowledge Workers and Meritocracy

This pattern is straightforward: AI is a multiplier. Multiply ambition and you get more output. Multiply passivity and you get more comfortable irrelevance.

From a common-sense conservative standpoint, this is exactly how meritocracy is supposed to work. The people who put in mental effort, who take ownership of their own development rather than waiting for a system to manage it for them, are going to pull ahead. That's the correct outcome.

What government or employers do with this information is a separate question. If companies start benchmarking AI-assisted output at 2026 levels and calling that the floor, workers who don't adapt face real wage and advancement pressure. That's a workforce policy question that neither political party has seriously engaged yet.

The Open Question

ActivTrak's 10,000-worker dataset is large, but it captures early adopters — a self-selected group already disposed toward new tools. Whether the same intensity patterns hold as AI becomes ambient and mandatory rather than optional is genuinely unknown. The workers who are coasting now may be forced to engage when their employers set AI-augmented baselines as standard. Or the cognitive miser cohort may find ways to delegate even the oversight. The research doesn't resolve that yet.