How the latest AI technology can lead to a nation of shorter hours, without reductions in livelihoods or productivity.
This paper offers an equitable route for the deployment of AI in the pursuit of greater productivity. It builds upon leading analyses of the potential impacts of new AI technologies (specifically Large Language Models). It analyzes the potential eligibility for a four-day work week across U.S. states, driven by AI-enhanced productivity gains over the next decade.
Two scenarios are detailed and mapped: The first, where the productivity gains of AI are captured so as to deliver 20% reductions in working hours, whilst keeping pay the same for workers. The second considers how many workers could be augmented by AI to the extent that their productivity improves by at least 10%, thereby enabling a reduction of 10% of the working weekly working schedule, whilst keeping pay the same.
The paper finds that, by 2033:
- A four-day week (32-hours as full-time equivalent) could be achieved within a decade for 35 million workers (28% of the US workforce)
- 71% of the US labour market (128 million workers) could have working hours reduced by at least 10%, should Large Language Models be introduced into workplaces and used as the basis for increased free time.
- States with the highest proportion of workers that could work four-day weeks within the next decade include: the District of Columbia, Massachusetts, Utah, Washington, Maryland, Colorado, Illinois, New Hampshire, Virginia, Connecticut. In all >25% of the workforce could move to a four-day week with AI augmentation.
- The report thus recommends that public and private sector employers take advantage of this significant opportunity, to become world leaders in the take up of workplace AI as well as improve the lives of well over 170 million people.