Tech giants are gaga over AI, but employees say the AI race is making their jobs harder
AI companies say their tools help employees accomplish dramatically more, but workers describe punishing schedules. The emerging workplace battle may ultimately be about who gets to keep the time AI saves.
Tech companies are investing heavily in AI, but some employees claim the AI race is actually making their jobs harder. At OpenAI, Anthropic, Meta, and Google, workers have reported punishing schedules and AI projects that consume evenings and weekends. Despite the promise that AI will help people accomplish more in less time, productivity gains are not translating into shorter workdays.
OpenAI's Codex, for instance, has become the go-to AI work tool across departments, with employees delegating longer and more complex tasks to it. Anthropic's internal data shows that more than 80% of its codebase is now authored by its Claude model, with engineers merging eight times as much code per day in 2026 compared to 2024.
However, Anthropic notes that code volume might slightly overstate productivity gains. Employees at OpenAI and Anthropic have reported working 70 hours a week, with some even working through weekends to catch up. At Meta, workers have been assigned to urgent AI projects that push them into nights and weekends. Google employee Amin Shali linked his increased workload to internal engineering problems as resources were redirected towards AI projects, resulting in improved sleep and health after leaving the company.
UC Berkeley research suggests that generative AI use can make employees faster and expand their task attempts, often occurring during periods that were previously considered breaks. OpenAI proposes that when AI reduces routine workloads, companies should consider 32-hour or four-day workweek trials, potentially leading to shorter weeks or paid time off.
South Korea plans to make AI access as accessible as electricity or public transit, paying for premium tools to boost domestic AI chatbot usage and reduce dependence on US and Chinese models. Meanwhile, AI-powered cyberattacks are a growing concern, with a group warning that they could become more widespread and sophisticated, posing risks to hospitals, water treatment plants, and internet infrastructure.
Anthropic has launched a Model Hardware Standard (MHS) to safely control lab instruments and factory equipment with AI, reducing integration time from months to hours or minutes.
Written by urgent.news from Digital Trends's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
