Who Will Control AGI? Regulation and Governance Explained

Short answer: Right now, nobody fully controls AGI. Control is split between a handful of AI companies, national governments (mainly the United States, the European Union and China), the companies that make advanced chips, and new international bodies at the UN. The US is favoring innovation and limiting state rules, the EU is enforcing the AI Act, and China is building a state-led model while promoting its own global organization. No law anywhere yet regulates "AGI" by name.

Every powerful technology eventually forces the same question: who holds the keys?

With nuclear energy, the answer became governments, treaties and inspectors. With the internet, the answer became a messy mix of companies, standards bodies and national laws. With AGI, we are deciding the answer right now, mostly without the public noticing.

I spend a lot of time with government leaders as well as business leaders. One of my keynotes is called Fearless Government for a reason. The public sector is being asked to govern a technology that moves faster than any law ever written. That is not a reason to give up. It is a reason to understand the landscape clearly.

This article explains who is shaping AGI's rules, how the major approaches differ and what it means for your organization. If you are new to AI governance inside a company, start with What Is AI Governance and Why Does It Matter for Executives?

Why Controlling AGI Is So Hard

Before we look at the players, it helps to understand why this is one of the hardest governance problems in history.

  • Nobody agrees on the definition. Regulators cannot easily govern something they cannot define. That is why most laws regulate "general-purpose AI" or "frontier models" based on capability and risk, not an AGI label. I explain the definition debate in How Will We Know When AGI Arrives?

  • It moves faster than law. Writing, passing and enforcing a law takes years. AI capability changes in months.

  • It crosses borders. A model trained in one country can be used everywhere within minutes.

  • It is concentrated. Fewer than ten organizations can build frontier AI. That gives a few players enormous influence. I map them in Who Is Building AGI?

  • Some of it cannot be recalled. Once an open model's weights are released publicly, they cannot be taken back.

The Seven Forces That Control AGI Today

1. The AI companies themselves

For now, the most direct control sits with the organizations building frontier AI. Their boards, leaders and internal safety teams decide what to build, how to test it and when to release it.

Many publish voluntary safety frameworks. According to the International AI Safety Report 2026, 12 companies published or updated frontier AI safety frameworks in 2025. Critics point out the obvious problem: these are mostly self-written and self-enforced.

There is also a corporate governance twist. At least one major AI company has stated that its own board decides when AGI has been reached, and AGI declarations have been written into major commercial contracts. One of the industry's largest partnerships now requires any AGI declaration to be verified by an independent expert panel. That tells you how much money is riding on the word.

2. The United States

The US approach in 2025 and 2026 has been to prioritize innovation and competitiveness.

  • In January 2025, the administration rescinded earlier AI safety directives and issued a new executive order focused on removing barriers to AI leadership.

  • In July 2025, it released an AI Action Plan centered on accelerating AI development and infrastructure.

  • In December 2025, it signed an executive order titled "Ensuring a National Policy Framework for Artificial Intelligence." It created a Justice Department AI Litigation Task Force, which began work in January 2026, to challenge state AI laws the administration sees as too burdensome.

But executive orders cannot by themselves cancel state laws. As of November 2026, Congress has not passed a law that broadly overrides state AI rules. In 2025, the Senate voted 99 to 1 to strip a proposed moratorium on state AI laws from a major bill, and a bipartisan draft federal framework released in June 2026 has not advanced.

Meanwhile, states are acting. California's Transparency in Frontier Artificial Intelligence Act took effect on January 1, 2026. It requires the largest AI developers to publish frontier AI frameworks explaining how they manage catastrophic risks, report serious safety incidents and protect whistleblowers, with penalties of up to $1 million per violation.

The simple version: in the US, the federal government is pushing for speed and one national standard, while states are writing their own rules, and the courts and Congress will decide who wins.

3. The European Union

The EU has the world's most comprehensive AI law: the AI Act. It regulates AI by risk level.

  • February 2025: bans on unacceptable AI practices and AI literacy requirements took effect.

  • August 2025: obligations for providers of general-purpose AI models began, including transparency and extra duties for the most powerful models that pose "systemic risk."

  • August 2026: the EU's AI Office gained full enforcement powers. Fines for general-purpose AI violations can reach €15 million or 3% of global annual turnover.

  • July 2026: a "Digital Omnibus" amendment entered into force. It delayed rules for high-risk AI systems to December 2027 for standalone systems and August 2028 for systems built into regulated products.

Here is the key point many headlines missed. The delay did not apply to the rules for powerful general-purpose models. Those are already in force. For anything close to AGI, the EU's rules apply now.

The simple version: the EU regulates by risk, has real fines, and its general-purpose AI rules are already live, even though some other deadlines slipped.

4. China

China's model is state-led and built around registration and content control.

  • Algorithms that shape what people see must be registered with regulators (since 2022).

  • Rules cover deepfakes (2023), generative AI services (2023) and labeling of AI-generated content (2025).

  • In July 2025, China published a Global AI Governance Action Plan with a 13-point roadmap.

  • In July 2026, China announced the establishment of the World Artificial Intelligence Cooperation Organization, headquartered in Shanghai, with 29 countries reported as founding participants.

The simple version: China controls AI tightly at home and is actively building international alliances to shape global rules, especially with developing countries.

5. The United Nations and global bodies

Global coordination is still early, but it is real.

  • In August 2025, the UN General Assembly created an Independent International Scientific Panel on AI. Its 40 members were appointed in early 2026.

  • The panel released its preliminary report in July 2026, presented at the first UN Global Dialogue on AI Governance in Geneva on July 6 and 7, 2026. A second dialogue is planned for New York in 2027.

  • Separately, the International AI Safety Report, written by more than 100 experts and backed by over 30 countries, gives governments a shared scientific picture of AI risks.

These bodies cannot force anyone to do anything. Their power is shared evidence and diplomatic pressure. Some experts argue AGI will eventually need something stronger, such as an international agency with inspection powers, similar to nuclear oversight. Nothing like that exists yet.

6. The chip chokepoint

This is the quietest form of control, and one of the most powerful. Building frontier AI requires advanced chips made by a very small number of companies. The US has used export controls on advanced chips and chipmaking equipment since 2022 to slow rivals' access. Some allies have joined those restrictions.

Whoever controls chips, data centers and energy has real leverage over who can build AGI at all. But this lever is weakening. Chinese developers have narrowed the performance gap through efficiency and domestic chips.

7. Investors, customers and the public

Do not underestimate this one. Investors decide which labs get funded. Enterprise customers decide which AI they trust with their data. Employees decide where to work. Public opinion shapes what politicians do. Every time a large organization asks a vendor, "How do you test this for safety?", it nudges the industry.

AGI Governance at a Glance

1. Philosophy

  • United States: Innovation first

  • European Union: Risk-based rules

  • China: State-led control

  • International (UN): Shared science and dialogue

2. Main tools

  • United States: Executive orders, export controls, state laws

  • European Union: AI Act, AI Office, fines

  • China: Registration, content rules, labeling

  • International (UN): Scientific panel, Global Dialogue

3. Rules for powerful models

  • United States: State level (e.g., California), no broad federal law

  • European Union: General-purpose AI obligations in force since 2025

  • China: Generative AI rules and registration

  • International (UN): Non-binding

4. Big open question

  • United States: Federal vs state power

  • European Union: Can rules keep pace with capability?

  • China: Global influence vs openness

  • International (UN): Can anyone enforce anything?

The Big Governance Gaps

As of late 2026, here is what is still missing.

  1. No AGI-specific rules. Laws cover general-purpose and frontier AI, but none were written with genuinely human-level AI in mind.

  2. No global referee. There is no agreed body that could certify AGI, inspect labs or enforce safety across borders.

  3. Falling transparency. Stanford's Foundation Model Transparency Index fell from an average of 58 to 40 in its latest report. Governments mostly know what companies choose to tell them.

  4. Speed mismatch. Human oversight works at the pace of meetings. AI works at machine speed.

  5. Power concentration. If AGI arrives through a handful of organizations, the governance question becomes a question about democracy itself.

What This Means for Your Organization

You do not need to solve global AGI governance. But you do need to operate inside it. Here is what I recommend.

  1. Map your exposure. List every country and US state where you build, sell or use AI. Rules differ, and some already apply.

  2. Treat EU rules as your baseline if you operate there. They are the most detailed and are already being enforced.

  3. Ask vendors for their safety frameworks. Frontier developers are increasingly required to publish them. Read them. Use them in procurement.

  4. Make AI governance a board topic. Regulation, liability and reputation risk all land at the top.

  5. Plan for change. US federal and state rules could shift quickly. Build compliance processes that can adapt, not ones tied to a single law.

For the full readiness plan, see How Should Companies Prepare for AGI? Public sector leaders may also find Government Innovation in the Age of AI useful. And for a quick take on AI in the public sector, watch my video Status in Governments: Generative AI in the Public Sector.

My Take: Governance Is a Leadership Job, Not Just a Legal One

The way most organizations treat AI governance today is as a compliance checklist. I think that is a mistake.

Governance is really about one question: who is allowed to make which decisions, and who is accountable when something goes wrong? That is a leadership question. It is the same question nations are wrestling with about AGI, just at a smaller scale.

The organizations that answer it clearly inside their own walls will be far better prepared, whatever the rest of the world decides. And trust, built through clear accountability, will become one of the most valuable assets any company has in an AI world.

Conclusion: Control Is Being Decided Right Now

So, who will control AGI? Today the honest answer is: a small group of companies, a few powerful governments, the makers of advanced chips and an early, still-weak international system. The US is betting on speed. The EU is betting on rules. China is betting on state coordination and global alliances. The UN is betting on shared science.

None of them has the full answer yet. The rules being written in 2026 and 2027 will shape who benefits from AGI, who bears its risks and how much ordinary people get a say. That makes this one of the most important stories of the decade, and one every leader should follow.

If your leadership team, association or public sector audience needs a clear, balanced conversation about AI and governance, explore my AI keynotes or book me for your next event.

Frequently Asked Questions About AGI Regulation and Governance

Who will control AGI?

No single entity controls AGI today. Control is shared between the companies building frontier AI, national governments (especially the US, EU and China), the makers of advanced AI chips, and international bodies such as the UN's scientific panel on AI. How that balance develops is still being decided.

Is AGI regulated?

Not by name. No law yet regulates "AGI" specifically. However, powerful general-purpose AI is regulated in several places, including under the EU AI Act, California's frontier AI transparency law and China's generative AI rules.

How does the EU AI Act affect AGI?

The EU AI Act sets obligations for providers of general-purpose AI models, with extra requirements for the most powerful models that pose systemic risk. These obligations have applied since August 2025, and the EU's AI Office gained full enforcement powers in August 2026. A 2026 amendment delayed some high-risk rules but not the general-purpose AI rules.

What is the US doing to regulate AI?

The US federal government is prioritizing AI innovation. A December 2025 executive order created a Justice Department task force to challenge state AI laws seen as burdensome. Congress has not passed broad federal AI legislation as of November 2026, while states like California have passed their own laws.

Is there a global AI regulator?

No. The UN created an Independent International Scientific Panel on AI in 2025 and held its first Global Dialogue on AI Governance in July 2026, but these bodies cannot enforce rules. Some experts propose an international agency for advanced AI similar to nuclear oversight, but none exists.

Who decides when AGI has been achieved?

There is no official referee. Some AI companies have said their boards decide, and at least one major industry partnership now requires an independent expert panel to verify any AGI declaration because significant commercial rights depend on it.

What should businesses do about AI regulation?

Map where your AI is built, sold and used, treat the EU AI Act as a baseline if you operate in Europe, ask AI vendors for their published safety frameworks, put AI governance on the board agenda and build compliance processes that can adapt as rules change.

About Shawn Kanungo

Shawn Kanungo is a globally recognized innovation strategist, keynote speaker and bestselling author of The Bold Ones. He spent 12 years advising leaders on disruptive innovation at a global professional services firm and has more than two decades of experience working with Fortune 500 companies, associations and governments. He helps business and public sector leaders turn disruption into opportunity. Watch more at watch.shawnkanungo.com.

Previous
Previous

AGI Hype vs Reality: How to Read the Headlines

Next
Next

How Should Companies Prepare for AGI?