Overview
- BBC interviews relayed by multiple outlets on Tuesday collected accounts from current and former employees at OpenAI, Anthropic, Meta and Google who describe regular 50–90+ hour weeks, prolonged multiweek sprints and weekend crisis meetings.
- Several companies have publicly promoted four‑day or shorter workweeks based on AI productivity, but a former OpenAI technical employee told the BBC the firm did not trial a four‑day week internally while they worked there.
- Academic and industry studies from Harvard Business School, UC Berkeley and the AI Work Institute find that AI speeds individual tasks yet expands work scope because staff must take on more assignments and spend time checking or fixing AI outputs.
- Workers report managers reallocating staff to AI projects with little notice, executives modeling extreme hours, and performance regimes that raise pressure and increase burnout and attrition among engineers and product teams.
- The pattern suggests that near‑term gains from AI are being converted into faster product cycles and higher expectations rather than shorter schedules, a shift that could prompt new labor rules, company trials, or retention challenges to watch for.