Artificial intelligence was supposed to save workers time. Instead, many employees are discovering that faster work often just leads to more work.
A new ongoing study from the University of California, Berkeley found that AI tools inside one U.S. tech company did not reduce workloads. Researchers embedded themselves inside the 200-person company for eight months and observed how employees used enterprise AI tools in daily work. What they found was less task replacement and more about intensification.
As workers became faster with AI, expectations expanded around them. Product managers started taking on coding tasks. Designers moved deeper into engineering work. Software engineers spent more time reviewing AI-assisted output generated by coworkers. Employees multitasked constantly, running several AI prompts at once while juggling meetings, emails, and deadlines.
Even breaks began shrinking. Workers reportedly squeezed in “one more prompt” before lunch, after hours, or between meetings.
The study indicates something larger happening across the workforce right now: AI is becoming less of an optional tool and more of a workplace enhancement workers feel pressure to use just to keep up.
Productivity Gains Are Becoming Productivity Expectations
For years, workplace technology has been sold as a way to reduce friction and free up time. But historically, efficiency gains rarely stay personal for long. Once companies realize employees can produce more in less time, the baseline expectation often changes.
The same dynamic now appears to be unfolding with AI.
Instead of reducing workloads, AI is allowing organizations to compress more responsibilities into fewer roles. Tasks that once required additional headcount are increasingly being folded into existing jobs under the assumption that AI can help workers manage the load.
That pressure is already appearing in labor data.
Roughly 13% of U.S. workers now use AI every day at work. At the same time, more than 50,000 U.S. job cuts in 2025 were reportedly tied to AI-related restructuring, even as many organizations still lack systems capable of fully replacing human workers.
In practice, many employees are not being replaced outright, but they are being asked to produce more.
The New Competitive Layer at Work
AI use is also creating a new type of competition between workers; employees who use AI well can often complete tasks faster, respond quicker, and manage larger workloads. That changes workplace expectations for everyone else, including workers who may not want to rely heavily on AI tools or who are still learning how to use them effectively.
In some industries, using AI is already starting to resemble earlier workplace changes brought on by smartphones, email, or always-on messaging platforms. What begins as an advantage quickly becomes normalized.
Workers who do not participate can begin falling behind on speed, output, or responsiveness. That does not necessarily mean employees are working better. It often means they are working continuously.
More Output, Same Human Limits
The Berkeley researchers noted that while AI accelerated many tasks, the surrounding work systems largely stayed the same. Meetings still happened. Approvals still existed. Deadlines still stacked up.
The result was not less work, but denser workdays.
That tension is becoming part of a larger future-of-work conversation around burnout, attention, and workload expansion. As AI lowers the time needed to complete individual tasks, organizations may continue increasing expectations faster than workers gain relief.
Companies are now entering a phase where AI strategy is about deciding what happens after productivity rises. Once AI becomes part of everyone’s workflow, higher output quickly stops feeling exceptional and starts becoming the expectation.
