AI Work Hours Rise as Tech Employees Report Longer Weeks
AI work hours are rising for some technology workers despite promises of greater efficiency. Employees at major firms report longer weeks and intense project demands.
Tech leaders have predicted that AI will reduce working hours. Some executives even suggested that employees could eventually move toward four-day work weeks. However, workers describe a different reality. Some employees at leading AI companies report working between 70 and 90 hours during intense periods.
AI Projects Create Longer Weeks
A former OpenAI employee told the BBC that they often worked at least 70 hours weekly. They described weekend work, crisis meetings, and intense performance reviews.
Other workers described similar conditions at OpenAI and Anthropic. Their project sprints can reportedly stretch across several weeks. Meta employees also reported longer hours on urgent AI projects. Some workers said managers moved them onto these teams without giving them much choice.
These employees described late nights, weekend work, and constant availability. Meta has reportedly reduced some of the pressure from its earlier approach. Meanwhile, AI workloads can affect employees outside AI development teams. A former Google employee said resource shifts toward AI projects created additional engineering problems.
Those issues sometimes forced engineers to work late into the night. He later said his sleep and overall health improved after leaving Google. Research from UC Berkeley also challenges the idea of automatic time savings. Researchers followed technology workers for eight months while they used AI tools.
The study found that employees worked faster and handled broader responsibilities. They also extended their work across more hours of the day.
MIT researcher Neil Thompson explained why these patterns can continue. Companies often use saved time for new tasks, improvements, and checks.
Workers must also review AI-generated results for accuracy. Therefore, faster tools can create additional responsibilities instead of reducing workloads.
The evidence suggests that AI may change work rather than simply shorten it. Unless companies limit workloads, efficiency gains could lead to even more demands.
