Artificial intelligence was supposed to give workers something valuable: time. As AI systems become better at writing, coding, research, analysis, and routine office tasks, many technology leaders have argued that people should eventually work fewer days while keeping productivity high.
Yet the experience of many tech workers tells a different story. Instead of using productivity gains to create more free time, companies can use them to raise expectations. The result is a workplace where employees finish tasks faster, only to receive more tasks.
This creates a difficult mismatch between the public message and the daily experience of employees. Leaders may discuss freedom, flexibility, and efficiency, while teams are judged against faster deadlines and larger targets. The technology itself is not choosing between those outcomes. Company policies, incentives, staffing decisions, and workplace culture determine how its gains are distributed instead.
AI can increase the amount of work
A recent UC Berkeley Haas study offers an important explanation for this contradiction. Researchers observed workers at a technology company for eight months and found that generative AI often intensified work instead of reducing it.
That creates a subtle problem. A worker may feel more productive because an AI tool saves time on one activity. But the saved time does not necessarily become leisure. It can become another meeting, another feature, another document, or another deadline.
AI can also create extra checking work. Generated code, reports, summaries, and other material still need human review. When the output is unreliable or incomplete, employees must spend time finding mistakes and correcting them.
The contrast becomes especially striking in the technology industry itself. Companies building advanced AI products operate in an extremely competitive environment. New models, features, and products can become outdated quickly, creating strong pressure to move faster.
Reports about employees at companies such as OpenAI and Meta have highlighted periods of intense work, including sprints that can stretch to extremely long weeks. Some workers have reportedly faced schedules approaching 90 hours during demanding periods.
That is very different from the vision of AI as a tool for a better work-life balance.
The hidden cost of always being available
AI also removes some natural stopping points in the workday. Now, an employee can open an AI tool during lunch, continue a project in the evening, or run several processes while handling other work. The boundary between working time and personal time becomes easier to cross.
Over time, this can create fatigue even when individual tasks feel easier. A faster workplace is not automatically a healthier workplace.
What a shorter workweek would really require
AI could still support shorter working hours, but technology alone will not make that happen. Companies would need to decide that productivity gains should be shared with employees rather than used only to increase output.
Leaders also need to recognize that long hours can damage the very productivity they are trying to improve. Exhausted employees make mistakes, lose focus, and have less capacity for creative thinking.
A choice between efficiency and exhaustion
The debate around AI and working hours is more complicated than the promise of “work less with AI.” The technology can genuinely make people faster and more capable. Research also shows that it can expand responsibilities and push work beyond normal boundaries.
The future of work will depend on what companies do with that extra capacity. If AI becomes a reason to expect more work from fewer people, employees may simply become busier. If productivity gains are converted into shorter schedules, better flexibility, and more meaningful work, the technology could deliver the freedom it was supposed to create.
