By Kiran Pillai
Artificial intelligence is often discussed as a technology that will make workers faster. A more radical question is what happens when workers no longer need to spend as much time working.
If AI can perform a growing share of administrative, analytical, technical and creative tasks, society could use the resulting productivity gains to increase profits. But it could also use them to increase human freedom.
The four-day working week offers a practical starting point. Instead of allowing every productivity improvement to become another demand for higher output, companies and governments could convert a portion of AI-generated productivity into reduced working time without reducing incomes.
This would change the political meaning of automation.
The Old Promise of Technology
Industrial technology has repeatedly increased the amount that society can produce. Yet higher productivity has rarely translated automatically into proportional increases in free time.
Workers continue to sell their time because employment remains the primary mechanism through which most people obtain income.
AI creates a new challenge to this arrangement. Unlike earlier machines that primarily replaced physical effort, AI increasingly performs tasks involving language, calculation, analysis, coding and decision support. Contemporary socialist discussions are therefore increasingly focused on the question of who controls AI and how its productivity gains should be distributed.
The technology creates a possibility that was previously difficult to achieve at scale: producing more while requiring less human labour. The question is whether society will use that possibility.
Productivity Should Buy Time
Imagine an organization where AI increases productivity by 25 percent. There are several possible outcomes.
Management could demand 25 percent more output. The company could reduce its workforce. Shareholders could receive higher profits. Or employees could work fewer hours. The fourth option deserves much greater attention.
If technology makes production more efficient, workers should not necessarily be required to produce continuously more goods and services. Part of the gain could become additional leisure.
A five-day working week could become four days. Later, it could become three and a half.
The objective would not be to eliminate work. It would be to eliminate unnecessary work.
AI Should Reduce Drudgery
Many jobs contain repetitive activities that consume enormous amounts of human attention. Employees spend hours preparing reports, entering information, searching documents, scheduling meetings, answering routine questions and processing administrative requests. AI can automate some of these activities.
But the resulting time should not automatically be filled with additional tasks. If an employee previously spent eight hours completing work that AI can help complete in six, the social achievement is not necessarily that the employee should complete twelve hours of work.
The achievement may be that the employee gets two hours back. That distinction is fundamental.
The Four-Day Week as an AI Dividend
A four-day working week could become a practical form of the AI dividend. Instead of distributing AI productivity entirely through wages or profits, society could distribute some of it through time.
Workers would receive something that cannot easily be measured in conventional economic statistics: more life.
More time with children. More time for education. More time for community activities. More time for exercise, culture and creativity. More time for rest.
Economic policy usually treats income as the principal measure of welfare. AI could force us to reconsider the importance of time.
The Global South
This question is especially important for countries such as India. A large part of India’s workforce operates under intense time pressure. Millions work in offices, factories, shops, farms and informal businesses. AI could increase productivity across these sectors.
But productivity gains should not simply become an opportunity for organizations to demand longer availability from workers. India could instead experiment with AI-assisted reductions in working time.
Public institutions could test four-day schedules. Technology companies could link automation to working-hour reductions. Manufacturing cooperatives could distribute productivity gains between wages, investment and leisure.
The country could develop an economic model in which technological advancement gradually reduces the amount of time people need to spend performing necessary work.
What About Employment?
Critics may argue that shorter working hours will increase labour costs. But this assumes that employment must remain organized around maximizing hours. If productivity rises, the same economic output can potentially be produced with fewer labour hours.
The transition does require careful management. Some industries operate continuously and cannot simply close for three days. Hospitals, transport systems, manufacturing facilities and emergency services need shift-based arrangements.
But shorter individual working hours do not necessarily mean shorter operating hours. More flexible shift structures could distribute available work among more people.
Workers Must Have a Voice
The transition should not be imposed entirely by management. Workers need a role in determining how AI changes their jobs.
This is where trade unions and worker organizations become important. They can negotiate agreements covering AI deployment, retraining, surveillance, productivity measurement and working-time reductions.
The goal should be to transform technological change from something done to workers into something negotiated with workers. Recent socialist proposals have explicitly connected AI with workers’ control, shorter working weeks and social ownership of technology.
Beyond the Four-Day Week
The four-day week should not be viewed as the final destination. It is a transitional institution.
If AI and robotics continue increasing productivity, societies could eventually move toward much shorter working weeks. The long-term question becomes: How much human labour is actually necessary to maintain a high standard of living?
If the answer keeps declining, insisting that everyone must work forty or fifty hours simply because that is how industrial society was organized, it would become increasingly irrational.
Technology would have changed. The economic institution would not have caught up.
A Different Definition of Progress
There is a tendency to measure technological progress through computing power, company valuations and productivity statistics.
But there is another measurement. How much unnecessary work has technology removed from human life?
A society in which AI makes people work harder is technologically advanced but socially primitive. A society in which AI allows people to work less while maintaining good living standards has achieved something much more meaningful.
This does not require abolishing markets or private enterprise. Companies can continue innovating. Entrepreneurs can continue building AI products. Competition can continue.
But the gains from innovation can be shared through wages, public services, worker ownership and time. The central political choice is therefore not whether AI will make humans work. It is what society will do with the work that AI makes unnecessary.
If the answer is simply higher profits, artificial intelligence may deepen existing inequalities. If the answer includes shorter working weeks, greater economic security and more time for human development, AI could become one of the most liberating technologies in history.
The ultimate measure of an intelligent society may not be how much its machines can produce. It may be how much unnecessary work its people no longer have to do.
Kiran Pillai is an Indian researcher and policy writer with a focus on economics, technology, governance and social development. He is pursuing doctoral research in business administration and writes on public policy, institutions, economic transformation and contemporary political economy.
Image: https://commons.wikimedia.org/wiki/File:Artificial_Intelligence_Word_Cloud.png Author: Madhav-Malhotra-003, licensed under the Creative Commons CC0 1.0 Universal Public Domain Dedication.
