Its employees are more productive with trimmed hours, and with the addition of AI tools such as ChatGPT earlier this year, the four-day week has felt even more seamless, Earthly CEO Oliver Bolton tells Fortune. Earthly now uses the platform to sift through projects, brainstorm, research and streamline operations overall, which has freed up more time for company staff.

“The consensus is when you’ve got four days to get your work done, it gives you that much more focus,” Bolton said. “I see AI as a great opportunity to just be more productive, work more efficiently, get more done to a high level of quality. We’ve had the 4-day workweek without any AI for over a year, so we’ve got that experience. With AI, it can enable us to do more.”

Soon, some of the benefits Earthly has experienced could be seen across Britain—AI could reduce the hours worked by at least 10% for a whopping 88% of its workforce, according to a recent report by Autonomy, which helped carry out the world’s largest four-day workweek pilot last year.

“This represents a huge opportunity for policymakers, trade unions and of course the millions of workers who are likely to be affected in some or another by these new AI technologies,” the authors of the Autonomy report wrote. 

Two AI scenarios

The think tank considered two scenarios—first, where productivity gains from AI cut down hours at work by 20%, and the other, where workers’ jobs are augmented by AI such that their productivity increases by at least 10%. In either case, the report notes that over the next 10 years, 8.8 million Brits could benefit from a 32-hour workweek without suffering a loss in pay. 

“What we’re really trying to do is to say, if we use this technology [AI], for this particular purpose—in this case, we’re saying if it was used to increase productivity… how could the benefits be distributed more equitably or inclusively,” Autonomy research director Will Stronge told Fortune. “That’s why these particular studies are of interest to us because we can start getting to grips with what a full optimization of the tech would do.”

The argument for ChatGPT and similar tools could usher in a shorter workweek by increasing productivity isn’t new. A June note by investment bank Jefferies pointed to a “broader acceptance” for a four-day workweek, thanks to AI making people quicker at their current jobs.

Academics agree with this, too—earlier this year, Christopher Pissarides, the Nobel Prize laureate and London School of Economic professor who specializes in labor economics and the impact of automation, said he was “optimistic” about AI’s role in improving productivity. 

“We could increase our well-being generally from work and we could take off more leisure. We could move to a four-day week easily,” he said during a Glasgow conference in April.   

Alright, there’s more time—so what?

AI tools could soon usher in an era of just four days at work, opening up a lot more time for people. But the big question remains what people choose to do with their new-found time that AI tools help unlock, said Carl-Benedikt Frey, an associate professor of AI & Work at the Oxford Internet Institute. 

In an influential 2013 paper that Frey co-authored, he predicted that automation could eliminate nearly half of all U.S. jobs. The recent generative AI wave which has put the likes of ChatGPT in the spotlight is different, he says. He told Fortune in September that it isn’t an automation tech yet as it still needs a human to prompt it and give it commands—but it can certainly make people better at low-stakes tasks.

Still, Frey argues, “Any productivity-enhancing technology, in principle, can enable you to work less. The question is whether empirically that is the case.” He pointed out that the productivity boost in the U.S. during the 20th century led to shorter, 40-hour weeks (it used to be over 70 hours in some industries not too long before that) which didn’t necessarily translate into an equivalent increase in leisure time. Similar results have been found in Britain as well. 

“We could have taken all that productivity gains out in leisure, but people decided to continue to work,” Frey said, adding that this could’ve been for a number of reasons including the preference for higher incomes by working more.

“So, it’s a question of choice, and those choices may differ depending on institutions in place, personal preferences and on a variety of [other] things.”

While it could be years before we see a sharp shift towards using our extra hours on leisure rather than work, Frey is already starting to see changes in worker preferences. And data reaffirms that, too—for instance, workers are willing to accept pay cuts just to be able to work 32-hour weeks instead of the usual 40-hour week, jobs board Indeed data in the U.K. reveals. 

The four-day workweek pilot in 2022, whose results were released in February, marked a major breakthrough with a 92% success rate among the U.K.’s 61 participating companies. Companies also saw improved job retention and mental and physical health of employees, who took fewer sick days and reported greater work-life balance.

More long-term advantages of a shorter workweek include greater gender equality, as it offers flexibility to employees when it comes to childcare responsibilities that tend to be borne by women, experts argue. 

With a groundswell of industry leaders and authorities calling for stronger AI regulations as it becomes more widely available to people, it can be hard to predict the tech’s trajectory. But one thing is certain: AI is quickly reshaping the world of work as we know it by lending more momentum to the shift to greater leisure. 

Earthly’s Bolton encourages the firm’s employees to use their time pursuing meaningful hobbies—so now, his employees use their time for wide-ranging activities from tending to chickens to mentoring startups and upskilling.    

There are clearly important up-sides that four-day weeks offer—but it hinges on AI being implemented fairly across the economy, Autonomy’s Stronge argues.   

“I think once GPT or [other] large language models in general become as ubiquitous as email, that’s when we’ll reach a new level or new plateau of productivity,” he said. “I think we’re not quite there yet.”