Who Controls AI, and Who Is Left Behind?

Artificial intelligence is often presented as a technology that belongs to everyone. Anyone with an internet connection can open an AI chatbot, generate an image, translate a document, or use an intelligent assistant to solve a problem. From the outside, this seems like a technological revolution in which everyone has an equal opportunity to participate.

But is that really the case?

Imagine two people entering the AI-powered world. One works for a technology company with access to powerful computers, enormous datasets, experienced engineers, and advanced AI systems. The other lives in a community where internet access is unreliable, digital education is limited, and the cost of modern technology is difficult to afford.

Both people are living in the age of artificial intelligence, but they are not experiencing the same opportunities.

This raises an important question: If AI is becoming one of the most influential technologies in human history, who actually controls it, and who gets left behind?

The answer involves more than technology companies. It includes governments, researchers, investors, businesses, educational institutions, and ordinary people whose lives are increasingly affected by automated decisions.

Understanding this distribution of power is essential because the future of AI will not depend exclusively on what machines can do. It will also depend on who develops them, who owns the infrastructure, who benefits economically, and who has a voice in deciding how they are used.

Learning Outcomes

By the end of this lecture, readers should be able to:

1. Understand who develops, owns, and influences modern AI systems.
2. Explain why control over computing infrastructure and data creates technological power.
3. Recognize how AI can widen existing economic and educational inequalities.
4. Understand the digital divide and its relationship with AI accessibility.
5. Examine the effects of AI on developing countries and smaller businesses.
6. Identify why transparency, accountability, and public participation matter.
7. Explore ways to make AI opportunities more accessible to different communities.

Who Actually Controls Artificial Intelligence?

When discussing control over AI, it is tempting to imagine a single company or organization making all the decisions. In reality, AI development is distributed across several groups, each controlling different parts of the technology.

Let’s examine the major players.

1. Technology Companies: The Owners of Major AI Platforms

Large technology companies play a significant role in developing and operating advanced AI systems.

They invest in research, employ specialized engineers, build AI models, operate cloud infrastructure, and provide tools that millions of people use.

Their influence comes from several sources:

• Financial resources to support expensive AI research.
• Access to large computing facilities.
• Ability to acquire and process substantial datasets.
• Experienced research and engineering teams.
• Established platforms that distribute AI products worldwide.

When a company controls the platform through which people access an AI service, it can influence important decisions about availability, pricing, features, usage restrictions, and product development.

However, platform ownership does not mean that one company controls all artificial intelligence. The AI ecosystem includes commercial organizations, open-source communities, universities, governments, and independent developers.

2. Governments: The Regulators of AI Development

Governments influence AI through laws, regulations, public funding, procurement policies, and national technology strategies.

For example, governments may establish rules concerning personal data, automated decision-making, copyright, cybersecurity, and the use of AI in public services.

They may also invest in research institutions and digital infrastructure to strengthen their countries’ technological capabilities.

This creates another important dimension of AI control: the ability to determine which applications are permitted, which risks must be addressed, and what responsibilities developers and organizations must follow.

Government influence is particularly important when AI is used in areas such as healthcare, education, public administration, transportation, and financial services.

3. Researchers and Universities: The People Expanding AI Knowledge

Not every important AI development happens inside a major technology company.

Universities, independent laboratories, and research organizations contribute to machine learning, robotics, natural language processing, computer vision, and AI safety.

Their work helps expand scientific understanding and sometimes provides foundations for technologies later developed into commercial products.

Academic researchers also examine questions that commercial developers may not prioritize, including algorithmic fairness, environmental costs, privacy, and the social consequences of automation.

Their influence is different from that of platform owners. They may not control widely used AI products, but they can influence the scientific direction and public understanding of the technology.

4. Investors and Infrastructure Providers: The Financial Foundation

Developing and operating advanced AI systems requires substantial resources.

High-performance processors, data centers, electricity, networking equipment, and cloud services all contribute to the infrastructure behind modern AI.

Organizations that finance, manufacture, or operate these resources can influence which AI projects receive funding and which developers can access the computing power they need.

This is an important distinction: controlling an AI application is not the same as controlling the infrastructure that makes the application possible.

Key takeaway: AI power is distributed across platform owners, governments, researchers, investors, and infrastructure providers. Understanding these relationships helps explain why access to AI is not equally distributed.

The Hidden Resource Behind AI Power — Computing Infrastructure

When people use an AI chatbot, they usually see a simple interface: a text box, a response, and perhaps a button to generate an image.

What they do not see is the enormous infrastructure operating behind that interface.

Advanced AI systems depend on several resources:

• Computing power: Specialized processors and computer systems used to train and operate AI models.
• Data: Information used to develop, train, test, and improve AI systems.
• Electricity: The energy required to operate data centers and computing equipment.
• Cloud infrastructure: Remote computing services that allow organizations to access powerful machines.
• Technical expertise: Researchers, engineers, and specialists who build and maintain these systems.

These resources are not distributed equally around the world.

A university in a wealthy country may have access to advanced computing facilities and research funding. Meanwhile, a university in a lower-income country may struggle to obtain the equipment and financial support needed for comparable research.

This creates an imbalance that goes beyond access to AI applications. It affects who can actually participate in developing the technology.

The World Bank’s Digital Progress and Trends Report 2025 identifies connectivity, computing capacity, locally relevant data, and digital skills as important foundations for AI development.

Why Does Infrastructure Ownership Matter?

Consider a small technology startup that wants to develop an AI-powered educational platform.

The company may have a talented development team and a useful idea, but it still needs computing resources, training data, and funding.

If these resources are expensive or difficult to access, the startup may have to depend on services provided by larger technology companies.

This does not automatically mean that the startup cannot succeed. Cloud computing and publicly available AI models can reduce some barriers. However, dependence on external infrastructure can limit its flexibility and increase its operating costs.

The question becomes bigger than simply asking who has the best AI model.

It becomes a question of who owns the essential resources that allow others to participate.

This is one reason researchers discuss decentralized and more accessible AI development. Organizations such as the Institute for Decentralized AI examine how concentrated control can be challenged through alternative approaches to AI development.

Who Is Being Left Behind?

The benefits of artificial intelligence are not automatically reaching everyone.

Some people have access to high-speed internet, modern devices, professional training, and advanced digital tools. Others face barriers that prevent them from participating in the same technological transformation.

These differences create what is commonly called the AI divide.

The AI divide is not simply about whether someone can open an AI application. It also concerns whether people, communities, and countries have the resources and opportunities to use AI productively.

Let’s examine some of the groups that may experience these disadvantages.

1. Developing Countries

Developing countries face several challenges when trying to participate in the AI economy.

Limited digital infrastructure, expensive computing resources, shortages of specialized talent, and insufficient investment can make it difficult to develop domestic AI capabilities.

According to UN Trade and Development, approximately 100 firms control much of the world’s AI research, patents, and computing power. Its reporting also highlights the limited representation of developing countries in major AI governance discussions. UN Trade and Development (UNCTAD)+1

This creates an important concern.

Countries that do not participate meaningfully in AI development may become dependent on technologies designed elsewhere.

They might use advanced AI applications, but have limited influence over how those systems are developed, what languages they support, and which local problems they prioritize.

For example, an AI-powered agricultural application developed for large commercial farms may not address the needs of small farmers working with different crops, weather conditions, soil types, and financial resources.

The technology may be advanced, but its usefulness depends on how well it understands the environment in which it is being used.

2. People Without Reliable Internet Access

Imagine a student who wants to use AI to understand difficult mathematics lessons.

An AI tutor could explain concepts, generate practice questions, and provide additional learning material. However, the student needs a suitable device, reliable internet access, and sufficient digital knowledge to use it.

Without these basic resources, the potential benefits remain inaccessible.

The problem is not a lack of intelligence or motivation. It is a lack of opportunity.

UN Trade and Development reported that approximately 65% of people in least developed countries remain offline. This illustrates how basic connectivity can determine who gets to participate in the AI economy. UN Trade and Development (UNCTAD)

3. Workers Facing Economic Disruption

Artificial intelligence is changing how certain tasks are performed.

Writing, translation, data processing, customer support, research assistance, and some administrative activities can now be supported or partially automated by AI systems.

For some workers, this creates opportunities to become more productive. For others, it introduces uncertainty about the future of their occupations.

Consider a person whose professional work involves preparing routine reports. An AI system may be capable of completing some of those tasks much faster.

But what happens to the worker?

The answer depends on several factors, including the employer’s decisions, the nature of the job, opportunities for retraining, and the availability of alternative employment.

It is important to distinguish between a task being automated and an entire occupation disappearing. Many jobs contain responsibilities that require human judgment, communication, physical activity, and accountability.

Nevertheless, workers who have limited access to training may struggle to adapt when their industries change.

This is why AI education and professional development should not be considered optional extras. They are important parts of preparing people for technological change.

4. Small Businesses and Independent Creators

Large companies often have dedicated technology departments, financial resources, and access to professional AI services.

Small businesses operate under different conditions.

A local business may want to introduce automated customer support, improve product descriptions, or analyze sales information. However, it may not have the budget or technical expertise required to implement advanced systems.

Independent writers, designers, educators, and creators can face similar challenges.

AI tools have lowered some barriers to content production, but access to professional-grade features, reliable infrastructure, and specialized knowledge can still vary significantly.

The important distinction is that having access to an AI tool does not necessarily mean having the resources to compete equally.

The Language Problem — Whose Knowledge Does AI Understand?

There is another important dimension of AI inequality that receives less attention: language and cultural representation.

Many AI systems are trained using enormous collections of information. The composition of those datasets influences what models learn, which languages they handle effectively, and how well they respond to different cultural contexts.

Suppose a student asks an AI assistant a question in English. The system provides a detailed explanation with examples and supporting information.

Now imagine another student asking the same question in a less widely represented regional language.

The second student may receive a shorter, less accurate, or less natural response.

This difference matters because language is not simply a communication tool. It is connected to education, identity, culture, and access to knowledge.

Why Local Languages Matter

Consider communities where people primarily communicate in languages that receive less attention in AI development.

If AI systems do not understand those languages effectively, their users may have fewer opportunities to benefit from AI-powered education, healthcare information, and digital services.

The issue is particularly important for developing countries with multiple regional languages.

For example, an AI educational system designed mainly around English-language material may not automatically provide equally effective learning experiences in Urdu, Punjabi, Sindhi, Pashto, or Balochi.

Improving language representation requires more than translating existing content. It also involves understanding local expressions, educational contexts, cultural references, and the different ways people communicate.

An AI system cannot serve everyone equally well if it does not adequately understand the people it is intended to serve.

Who Decides How AI Should Be Used?

So far, we have examined who develops AI and who has access to it. But another important question remains:

Who decides the rules?

AI systems can influence decisions in recruitment, banking, education, healthcare, transportation, and public services.

For example, an organization might use an automated system to help screen job applications. Another might use AI to identify unusual financial transactions.

These systems can make processes faster, but they can also introduce errors, discrimination, or decisions that are difficult for affected people to challenge.

Imagine an applicant being rejected by an automated recruitment system without receiving a clear explanation.

Who is responsible for that decision?

Is it the organization that purchased the system, the company that developed it, or the people who configured and deployed it?

These questions demonstrate why AI governance matters.

The Importance of Transparency

Transparency means making relevant information about an AI system understandable to the people who use it or are affected by it.

This can include explaining:

• Where AI is being used.
• What purpose the system serves.
• What information it relies on.
• What its known limitations are.
• How mistakes can be reported or challenged.
• Who is accountable for important decisions.

Transparency does not mean that every user must understand the technical architecture of an AI model.

It means people should have enough information to understand the system’s role and the consequences of relying on its output.

Why Human Accountability Still Matters

AI can assist with decisions, but organizations should not use automation as an excuse to avoid responsibility.

A hospital cannot simply blame an algorithm when an AI-assisted process produces a harmful recommendation. Similarly, a financial institution should have appropriate procedures for reviewing automated decisions that affect customers.

Human oversight is particularly important when decisions involve people’s rights, safety, employment, finances, or access to essential services.

AI should support responsible decision-making, not make accountability disappear.

Can AI Make the Existing Inequality Even Worse?

Artificial intelligence has the potential to create opportunities, but it can also reinforce disadvantages that already exist.

To understand this, consider three possible situations.

The Productivity Advantage

A company with access to advanced AI tools can automate repetitive tasks, analyze information quickly, and develop products more efficiently.

A smaller competitor without comparable resources may struggle to maintain the same level of productivity.

Over time, this difference could influence market competition, business growth, and employment opportunities.

The Education Advantage

Students with access to AI-powered learning tools can receive additional explanations, practice exercises, and personalized learning support.

Students without suitable devices, connectivity, or digital literacy may not receive the same advantages.

Consequently, a technology that could help make education more accessible might also widen existing educational gaps if access remains unequal.

The Wealth Advantage

Organizations that own valuable AI infrastructure, intellectual property, and successful AI products may capture significant economic benefits.

Meanwhile, workers and smaller businesses may experience the costs of technological change without receiving an equivalent share of its rewards.

This does not mean inequality is inevitable. The outcome depends on investment, competition, education, public policy, and how organizations choose to distribute the benefits of technological progress.

The central concern is that technological progress alone does not guarantee social progress.

What Can Be Done to Ensure Nobody Is Left Behind?

If AI opportunities are unevenly distributed, what practical steps can help create a more inclusive technological future?

There is no single solution. Different communities require different investments and approaches.

1. Make Digital Education More Accessible

People need opportunities to understand AI before they can use it effectively.

Schools, colleges, universities, and professional training institutions can introduce practical AI education into their learning programs.

Students should learn how AI works, where it can be useful, what its limitations are, and how to evaluate its responses.

Importantly, AI education should not be restricted to computer science students. Business owners, teachers, writers, healthcare workers, and other professionals can also benefit from understanding these systems.

2. Improve Internet and Technology Infrastructure

Reliable internet access, affordable devices, electricity, and accessible computing services are essential foundations.

Without them, even the most advanced AI applications remain out of reach for many communities.

Governments, educational institutions, and private organizations can contribute by expanding connectivity and supporting affordable digital infrastructure.

3. Support Local AI Development

Countries should have opportunities to develop AI applications that address their own social, economic, and linguistic needs.

For example, local developers could work on:

• Agricultural assistance for small farmers.
• Educational applications supporting regional languages.
• Healthcare information tools adapted to local requirements.
• Digital services for small businesses.
• Systems that help preserve local languages and cultural knowledge.

Local development does not necessarily require every country to build the world’s largest AI model. Adapting existing technologies to local conditions can also create meaningful value.

4. Encourage Competition and Innovation

An AI ecosystem with multiple developers, research institutions, and technology providers can create more opportunities for experimentation.

Open-source models, accessible research, shared computing resources, and support for smaller technology companies can help reduce some barriers to participation.

Competition also gives organizations and individuals more choices instead of making them entirely dependent on a limited number of providers.

5. Protect Workers During Technological Change

Businesses and governments should consider how AI adoption affects employees.

Professional training, opportunities to develop new skills, workplace consultation, and responsible automation policies can help workers adapt.

The objective should not simply be to introduce AI as quickly as possible. It should also be to consider how productivity improvements can benefit people.

6. Include More Countries in AI Governance

AI is a global technology, and its consequences cross national borders.

Developing countries need meaningful opportunities to participate in international discussions about AI safety, transparency, access, and accountability.

If decisions about global AI development are shaped by only a limited group of countries and organizations, the needs of other communities may receive insufficient attention.

A more inclusive approach requires international cooperation and broader representation.

What Does This Mean for Ordinary People?

It is easy to read about AI companies, governments, computing infrastructure, and global inequality and assume that these issues concern only policymakers or technology professionals.

But AI governance and access affect ordinary people too.

Think about the choices you make when using an AI tool.

Are you relying on it to understand educational material? Are you using it to improve your professional skills? Are you allowing an automated system to influence an important decision?

These everyday interactions are part of a much larger technological transformation.

You do not need to become an AI engineer to participate in this future. However, developing certain abilities can help you navigate it more confidently.

Develop AI literacy: Understand the basic capabilities and limitations of artificial intelligence.

Strengthen critical thinking: Do not automatically accept an AI-generated answer as accurate.

Learn practical applications: Explore how AI can support your education, profession, or business.

Understand digital rights: Become aware of privacy, data protection, and the consequences of automated decisions.

Keep developing your skills: As technology changes, continuous learning can help you adapt to new opportunities.

Perhaps the most important point is that people should not be viewed merely as consumers of AI products. They should also have opportunities to understand, influence, and benefit from the technology shaping their lives.

The Future of AI — Will Power Become More Concentrated or More Distributed?

The future of artificial intelligence is not predetermined.

One possible direction involves increasing concentration. Advanced AI development could become increasingly dependent on organizations with access to expensive computing infrastructure, specialized talent, and substantial financial resources.

In this scenario, smaller companies and less technologically developed countries might rely heavily on systems developed elsewhere.

Another possibility involves broader participation.

More accessible AI models, affordable computing services, open research, better education, and locally developed applications could allow more organizations and communities to participate in the AI ecosystem.

Both possibilities deserve attention.

It is also possible for these developments to happen simultaneously. Some parts of AI development may become more concentrated, while other applications become easier for individuals and smaller organizations to access.

The important question is not simply whether AI will become more powerful.

It is whether the opportunities created by that power will become more widely available.

Three Questions That Will Shape the Future

As AI continues to develop, three questions deserve particular attention:

1. Will people have affordable access to the infrastructure and education needed to benefit from AI?
2. Will countries and communities have meaningful influence over the technologies that affect them?
3. Will the economic benefits of AI be shared broadly, or concentrated among a relatively small number of organizations?

The answers will help determine whether AI becomes a tool that expands opportunity or another factor reinforcing existing inequalities.

Frequently Asked Questions

Who controls artificial intelligence?

AI development and governance involve technology companies, governments, research institutions, investors, infrastructure providers, and other organizations. No single organization controls the entire AI ecosystem.

Why is AI controlled by a small number of companies?

Developing advanced AI systems requires significant investment, computing resources, technical expertise, and infrastructure. These requirements create barriers that can make it difficult for smaller organizations to compete at the technological frontier.

What does it mean to be left behind in the AI revolution?

It refers to individuals, communities, businesses, and countries that have limited access to AI technology, digital infrastructure, education, or opportunities to participate in AI development.

How does AI increase global inequality?

AI can reinforce existing inequalities when access to computing resources, education, investment, and economic opportunities is concentrated among already advantaged groups.

Are developing countries at risk of falling behind in AI?

Yes, particularly when they lack reliable connectivity, technical expertise, research funding, and computing infrastructure. However, adopting existing AI technologies and developing locally relevant applications can create opportunities for participation.

Can open-source AI reduce technological inequality?

Open-source and openly available models can reduce certain barriers by allowing more developers and organizations to study, adapt, and build upon existing technologies. However, access to computing resources, skilled people, data, and funding remains important.

Will AI replace human workers?

AI may automate or transform certain tasks, but its effects differ across occupations and industries. Some jobs may change, some tasks may disappear, and new responsibilities may emerge. Education and professional adaptation will be important.

Why is AI governance important?

AI governance establishes rules, responsibilities, and accountability for developing and using AI systems. It helps address issues involving safety, privacy, fairness, transparency, and public interest.

How can ordinary people prepare for an AI-driven future?

People can develop digital literacy, learn practical AI applications, strengthen critical thinking, understand privacy issues, and continuously improve their professional skills.

Conclusion: The Future Should Not Belong to Technology Owners Alone

Artificial intelligence is becoming an important part of how societies learn, work, communicate, and make decisions.

But technological progress should not be measured only by faster computers, more powerful models, or increasingly sophisticated applications.

It should also be measured by the opportunities it creates for people.

Who gets access to advanced education? Who can participate in technological innovation? Who benefits economically? Whose languages and cultures are represented? And who gets a voice when decisions about AI are made?

These questions matter because the consequences of AI will extend far beyond the companies developing it.

A student in a rural community, a small business owner, an independent researcher, and a worker adapting to automation all have a stake in how this technology develops.

The challenge is not to stop AI innovation. It is to make sure that innovation does not become an advantage reserved for those who already possess the greatest resources.

The real measure of an AI-driven future will not simply be how intelligent our machines become. It will be how many people are given the opportunity to benefit from that intelligence.