Left out OPINIONS

A.I. Public Utilities:

Making Intelligence a Basic Economic Infrastructure 

-By Kiran Pillai

October 11, 2026

Artificial intelligence is moving rapidly from an experimental technology into an economic infrastructure. It is entering offices, factories, hospitals, schools, financial institutions and government departments. The question is no longer whether society will use A.I.. The more important question is who will control the infrastructure on which that use depends. 

Electricity provides a useful comparison. Modern economies cannot function without reliable electricity, yet societies have developed different ownership and regulatory arrangements around power generation and distribution. Telecommunications, water and transport have followed similar paths. 

A.I. may require a comparable debate. 

If advanced intelligence becomes essential for education, production, research and public administration, treating it exclusively as a private subscription service could create a new form of inequality. Those who can afford the best models, computing resources and proprietary data will have greater productive capabilities than those who cannot. 

The alternative is to consider A.I. as a public utility. 

Intelligence as Infrastructure 

The distinctive feature of A.I. is that it can provide access to cognitive capabilities. 

A small business can use A.I. to analyze markets. A student can obtain tutoring. A local government can process large quantities of administrative information. A farmer can receive agricultural recommendations. A researcher can use A.I. to examine enormous datasets. 

This makes A.I. different from many ordinary consumer products. 

Its economic value increasingly comes from its ability to amplify the productive capacity of individuals and institutions. 

If access to that capability becomes essential, society has an interest in ensuring broad availability. 

Recent policy discussions are already moving in this direction. Public ownership and public participation in A.I. companies have become subjects of debate, including proposals for citizens to receive a financial stake in A.I.-generated wealth. 

A Public A.I. Layer 

A public A.I. utility would not mean that governments must own every A.I. company. Private companies could continue developing advanced models, applications and specialized services. The public sector could instead establish a basic layer of A.I. infrastructure. 

This might include publicly funded computing capacity, open models, public datasets, language resources and research infrastructure. Citizens could receive access to essential A.I. services without needing to purchase several expensive subscriptions. Universities and public research institutions could use the infrastructure for scientific work. Small businesses could access computational resources that would otherwise be beyond their budgets. 

The result would resemble public digital infrastructure rather than a traditional government monopoly. 

Why This Matters for the Global South 

The issue is especially important for developing economies. 

A country that depends entirely on foreign A.I. companies may become technologically dependent even while appearing digitally advanced. Its businesses may rent intelligence from overseas. Its government may rely on foreign cloud infrastructure. Its researchers may depend on external models. Its citizens may generate valuable data without participating in the economic value created from that data. 

A public A.I. infrastructure could provide another path. 

India, for example, could build A.I. systems specifically around Indian languages, agriculture, healthcare, education, manufacturing and public administration. 

The objective would not be technological isolation. It would be technological capacity. 

The Public A.I. Dividend 
Public infrastructure could also create a financial return. 

If governments invest in A.I. companies, computing infrastructure or strategic technology assets, part of the resulting returns could flow into a public fund.

That fund could finance education, healthcare, research and social protection. 

The concept is increasingly discussed as a form of social dividend from A.I.. Recent analysis has also warned that public ownership models can reproduce international inequalities if A.I. infrastructure depends on extracting resources from poorer countries. 

That warning is important. 

An A.I. dividend should not mean that one country captures the profits while another country provides the land, electricity, water and labor required to operate data centers. 

Public A.I. policy therefore needs an international dimension. 

Workers and Public A.I. 

A public utility approach should also involve workers. 

A.I. can increase productivity, but productivity gains do not automatically improve working conditions. A company may use A.I. to increase output while reducing employment, but a public-interest approach would establish different priorities. 

If A.I. allows a public institution to complete administrative work faster, the gain could be used to reduce employee workloads rather than simply eliminate positions. If automation improves factory productivity, workers could receive part of the benefit through shorter working hours, higher wages or profit-sharing. 

The underlying principle is simple: 

Technology should reduce unnecessary human effort rather than merely reduce the cost of human beings. Recent socialist analysis has similarly argued for worker control over A.I. and using productivity gains to shorten the working week. 

A.I. for Public Services 

The strongest case for public A.I. may be in basic services. 

Imagine a public A.I. assistant capable of explaining government schemes in every major Indian language. A citizen could ask about pensions, education scholarships, agricultural programs or business registrations using ordinary speech. The system could explain eligibility, identify required documents and guide the citizen through the process. 

A.I. could also help public hospitals manage appointments and records, help municipalities analyze infrastructure problems, and help schools provide individualized learning support. The objective would not be to replace teachers, doctors or government employees. It would be to give them better tools. 

Democratic Control 

Public ownership alone is not enough. A government-controlled A.I. system could become excessively centralized or intrusive. Citizens therefore need rights over how public A.I. operates. There should be transparency about important automated decisions. 

People should be able to challenge algorithmic decisions. Sensitive personal data should receive strong protection. Independent institutions should audit public A.I. systems. 

Workers should participate in decisions about workplace automation. The principle should be democratic control, not simply state control. 

An A.I. Commons 

The most ambitious model would combine public infrastructure with cooperative ownership. Universities could contribute research. Governments could contribute infrastructure. Workers could create cooperative A.I. systems. 

Small businesses could collectively finance specialized models. Communities could contribute datasets under agreed governance rules. Private companies could build commercial applications on top of the public layer. 

This would create an A.I. ecosystem containing markets, cooperatives, universities and public institutions. The result would be neither unrestricted corporate ownership nor complete state control. It would be a pluralistic technological economy with a strong public foundation. 

The Political Choice 

A.I. is often presented as an inevitable technological force. But technology does not determine its own economic structure. 

Society decides who owns infrastructure, how profits are distributed and what purposes technology serves. A.I. can become another mechanism for concentrating wealth. Or it can become an infrastructure through which knowledge and productive capacity become more widely accessible. 

The choice is ultimately political. If intelligence becomes one of the most important productive resources of the twenty-first century, access to intelligence may become as important as access to electricity, communication and education. 

That is why the debate should move beyond whether A.I. is good or bad. The deeper question is: 

Should access to artificial intelligence be treated as a privilege purchased from technology monopolies, or as a basic infrastructure through which society expands everyone's productive capabilities? 

The answer will shape not only the A.I. economy, but the economic structure of the decades ahead.