It was in 1930 that John Maynard Keynes envisioned a future where, thanks to technology, people in developed countries would work as little as 15 hours each week. By the time his prediction was supposed to have become reality, however, the shorter workweek had not materialized. Keynes was broadly right about rising living standards, but not about the shift toward dramatically more leisure.
Technologically, such a transformation was possible. But productivity gains did not automatically translate into fewer working hours.
As the late anthropologist David Graeber pointed out, technology has actually given rise to new kinds of work and bureaucracy. While automation has significantly reduced the amount of labor required in many manufacturing processes, Graeber argued that this did not lead to a corresponding reduction in working hours. Instead, there was a major expansion of administrative, managerial and other forms of white-collar work.
AI has yet to change the equation
Both Keynes and Graeber made their points way before artificial intelligence showed up and took over. Unfortunately, even in an era of mass digitization and widespread AI adoption, there is still no definitive answer to the question: Has technology increased productivity while reducing working hours? More troublingly, even some of the companies developing AI have yet to demonstrate that the technology is making work less demanding.
OpenAI, the company behind ChatGPT, has embraced the idea that AI-driven automation will increase productivity and eventually allow people to work fewer hours. The company has promoted discussions about a transition to a four-day workweek as AI makes workers more productive.
Yet OpenAI itself does not appear to be moving toward shorter working hours. On the contrary, reports suggest that late nights and weekend work have become common. According to one former OpenAI employee, the company had not even begun testing a four-day workweek.
The sprint culture
Technology companies commonly use the term «sprint» to describe periods of intense work before a major product launch. During these periods, employees may work long hours to meet tight deadlines.
At OpenAI, some employees have reportedly worked as many as 90 hours a week during particularly demanding periods — roughly 12 hours a day, seven days a week.
The situation is similar at Anthropic, another major AI company. Meanwhile, Meta CEO Mark Zuckerberg has cited AI-driven efficiency as one factor behind efforts to streamline the company and reduce its workforce. For employees who remain, however, greater efficiency does not necessarily mean less work.
When efficiency creates more work
According to innovation expert Neil C. Thompson, this is a common pattern in technology-driven workplaces. When an employee finds a way to automate or optimize a task and save time, that time is not necessarily returned to the employee. Instead, the worker may be given additional responsibilities or expected to produce more.
The same pattern is emerging in fields such as journalism and copywriting.
