As trials of the four-day working week gather momentum in the UK, a more radical idea is beginning to surface: could artificial intelligence make a three-day week possible?
The suggestion, made recently by investor and media executive Ari Emanuel, comes amid growing interest in how technology might reshape the structure of work.
Emanuel, who has raised nearly US$3 billion for a new live-events business, argued in an interview with the Financial Times that as AI becomes more deeply embedded in workplaces, it will save people time and increase disposable income. The result, he said, will be greater demand for leisure, culture and live experiences — the very areas his new venture, MARI, is designed to serve.
While his remarks are speculative, they raise a serious question for employers, HR leaders and policymakers: if AI can reduce time spent on work, could the standard working week shrink further, and should it?
Interest in four-day week continues to grow
The idea of a reduced working week is no longer theoretical. In 2022, 61 UK companies took part in the world’s largest trial of the four-day week, coordinated by the think tank Autonomy and the 4 Day Week Campaign. Nearly 3,000 workers were involved, with employers agreeing to cut hours by 20 percent without reducing pay. The results were overwhelmingly positive.
Participants reported lower stress and burnout, improved mental health and better work–life balance. Employers, for their part, noted stable or improved productivity, with some reporting revenue growth during the trial period. By the end of the six months, 92 percent of participating companies had decided to continue with the four-day model, and over half made the shift permanent.
The success of the pilot prompted further interest from both the public and private sectors. A new trial involving 17 UK companies began in 2024, with employees adopting either a four-day week or nine-day fortnight on full pay. The Scottish Government also ran its own pilot in two public bodies, with early results suggesting improved staff wellbeing and no loss in service delivery.
Efficiency gains from AI remain uneven
Against this backdrop, Emanuel’s suggestion that AI could accelerate the shift from four to three days of work has captured attention. Some early studies indicate that AI can deliver measurable time savings. According to research by the St Louis Federal Reserve, workers using generative AI tools reported saving an average of 5.4 percent of their weekly working time. In practical terms, that’s around two hours a week on a standard 40-hour schedule.
A separate experiment in the UK civil service, involving Microsoft’s AI-powered Copilot tools, found that employees gained back approximately 26 minutes a day, equivalent to nearly two weeks of time saved over the course of a year. Pearson has also projected that, by 2026, UK workers could save 19 million hours a week through automation of repetitive or low-value tasks.
But the figures don’t come close to allowing a full day, let alone two days, to be removed from the working week. In many roles, particularly frontline and service-based jobs, the scope for automation is limited. Tasks involving care, judgment, manual labour or human interaction are far less likely to benefit from time-saving AI tools.
From time saved to time off is not automatic
Even in highly automated sectors, there is no guarantee that time saved will be translated into time off. Studies of past productivity waves suggest that efficiency gains are often absorbed by rising workloads, not reduced hours. Without structural changes to job design, performance targets and organisational culture, AI may simply intensify work, compressing more into each day rather than reducing the number of days worked.
Work intensification is already a recognised risk in discussions around the four-day week. While many organisations in the UK pilot maintained productivity, some required staff to work at a faster pace or eliminate non-essential meetings and admin to fit their tasks into fewer days. Extending that model to three days would require much deeper redesign of work processes, job roles and expectations.
There is also the question of fairness. If time savings from AI are mostly available to desk-based professionals, while frontline workers continue on traditional schedules, the benefits of automation could widen existing inequalities. Any move toward shorter weeks would need to ensure that gains are shared across roles and pay grades, rather than reinforcing a two-tier workforce.
Practical barriers for employers
From an HR perspective, implementing a compressed working week presents numerous challenges. Pay, holiday entitlement, pension contributions and overtime would all need to be reconsidered. Legal requirements around working time, rest breaks and health and safety would also come into play, particularly for roles in regulated sectors.
Employers would need to invest in redesigning workflows, restructuring teams and training managers to lead in an outcome-focused environment, rather than one based on time spent at a desk. Communication and change management would be essential to ensure staff feel supported and that workloads remain sustainable.
Just as importantly, any shift would require cultural change. Leadership buy-in, trust in employees, and a willingness to judge performance on output rather than presence are all key enablers of successful four-day week models. Without them, the risks of burnout, resentment or informal overwork could outweigh the intended benefits.
Emanuel’s claim that AI could enable a three-day week is, at this stage, more provocation than proposal. It reflects a growing belief in the transformative potential of automation but also shows how much would need to change for such a model to be feasible or fair.
Still, the question is worth asking. As AI becomes more embedded in everyday work, organisations will face decisions not only about how to use it, but how to distribute the benefits. Whether that results in shorter weeks or simply different kinds of work, HR professionals will be central to shaping what comes next.
