- Published on
Tech leaders claim AI will usher in a four-day workweek, but employees still work 90 hours a week
- Authors
- Name
- DeepHorus
Major tech giants claim AI will bring about a four-day workweek, but a BBC investigation found that employees at these companies actually work far more than traditional standards, sometimes exceeding 90 hours per week. This stark contrast has drawn attention to the gap between AI promises and reality.
Google, OpenAI, Anthropic, and Meta invest hundreds of billions of dollars annually in AI research and development. Executives have long painted a future where AI takes over work and humans work fewer hours. Google's director of engineering once predicted that AI would achieve a four-day workweek by 2025, and OpenAI this year also publicly called on companies to trial a four-day workweek without pay cuts.
But reality is completely different. A former OpenAI employee told the BBC that they never saw a four-day workweek during their tenure; instead, they often faced crisis meetings, weekend overtime, and brutal performance reviews. They still had to go into the office on Saturdays or Sundays to catch up on progress or check systems, working at least 70 hours per week—far more than at other tech companies.
Similar situations have occurred at Anthropic and Meta. Anthropic claims Claude can work continuously for seven hours, and Zuckerberg claims AI changes the way we work, but Meta laid off 10% of its workforce this year. Anthropic's CEO warned that the more productive AI becomes, the more jobs it eliminates.
Tech industry insiders have revealed that "sprints" at OpenAI and Anthropic sometimes last for weeks, with weekly working hours exceeding 90 hours. A sprint is a high-intensity work period before launching a new feature, but at these companies it has evolved into a long-term state of overload.
Meta employees face the fate of being "conscripted." Current and former employees say that employees are transferred to AI teams without any advance notice, and they cannot refuse—otherwise, they have to resign. Once in an AI team, overtime becomes the norm, with late-night work and weekend shifts, and even after work they remain on standby.
Former Google employee Shali left in May this year. He revealed that AI projects caused frequent internal engineering system failures, and he often stayed up late solving problems. After leaving, his sleep and health improved significantly. He accused big tech companies of over-relying on AI and putting excessive pressure on engineers.
A study from UC Berkeley shows that AI has not reduced workload. Instead, it makes employees work at a faster pace, take on a wider range of tasks, and constantly check AI-generated content. MIT scholar Thompson pointed out that time saved will be swallowed up by new work, and the four-day workweek is nothing but a beautiful fantasy.
Shali lamented that AI should logically help people sleep better, but reality is the opposite. As AI further penetrates, how to balance efficiency and employee well-being has become a challenge that the tech industry must face. In the future, stricter labor regulations or corporate self-awareness may be needed to truly make AI benefit workers.
