Top AI Companies Are Asking for Supervision. Here’s Why Congress Might Not Give It to Them.

 

Top AI Companies Are Asking for Supervision. Here’s Why Congress Might Not Give It to Them.

**OpenAI, Anthropic, and Google are all saying the same thing: regulate us. They want independent watchdogs embedded in their labs, mandatory safety audits, and government oversight. But Washington is deadlocked, the President is calling AI safety a “hoax,” and the clock is running out. Here’s what’s really happening.**

---

## The Starting Gun

Let me paint the picture for you.

On Tuesday, September 15, 2026, OpenAI’s chief global affairs officer, Chris Lehane, did something that would have been unthinkable a few years ago. He threw his company’s weight behind a bipartisan House plan that would require AI developers to embed independent third-party evaluators inside their operations. Andrew Freedman, the co-founder of a nonprofit advocacy group called Fathom, called it a “starting gun”.

That same week, Anthropic CEO Dario Amodei published a 3,800-word essay calling for a coordinated slowdown of frontier AI development. He pledged that Anthropic would give “employee-like access to a team of embedded third-party evaluators” — watchdogs with the right to publish their findings without editorial control from the company.

Sam Altman agreed within hours. “I agree with Dario that we need to pace the frontier,” he posted on X. “Committing to having independent evaluators with employee-like access is a great idea, and we will do the same”.

Even Elon Musk backed the call. “Dario is right,” he wrote.

So the three most powerful AI companies in the world — companies with a combined market value that dwarfs most countries — are publicly asking the U.S. government to regulate them. And Congress is… not rushing to act.

---

## What the AI Companies Are Actually Asking For

Let me break down the specific proposals, because they’re more concrete than you might think.

### Embedded Third-Party Evaluators

This is the centerpiece. Both Anthropic and OpenAI have pledged to give independent evaluators **employee-level access** to their systems. Not just a conference room and a cup of coffee — actual desks, badges, and laptops. The evaluators would have the right to publish key findings about risk levels, incidents, and practices, without the company getting to edit their conclusions.

Anthropic’s proposal specifically named METR, a nonprofit that specializes in evaluating frontier AI models, as a potential partner.

### Mandatory Safety Audits

The FRONTIER Act — that’s Frontier Risk Oversight, National Transparency, Independent Evaluation, and Reporting — is the bipartisan House bill that would make these audits mandatory. Sponsored by Rep. Jay Obernolte (R-Calif.) and Rep. Lori Trahan (D-Mass.), it would require major AI developers to publish frameworks for assessing catastrophic risks and to disclose safety incidents.

It would also require cybersecurity measures to secure nonpublic model weights from unauthorized modification or transfer, and protocols for responding to critical safety incidents.

### A FINRA-Style Standards Body

This is the most ambitious idea. OpenAI, Anthropic, and Google have been discussing the creation of an industry standards body modeled on the **Financial Industry Regulatory Authority (FINRA)** — the private regulator that oversees U.S. brokers and investment firms.

The idea originated with Demis Hassabis, the co-founder of Google DeepMind, who proposed a **public-private partnership** overseen by the government but funded by the industry and staffed by independent technical experts.

Treasury Secretary Scott Bessent was reportedly considering a similar FINRA-like agency for AI that would report to the SEC.

### A “Duty of Care”

Senate Majority Leader John Thune, Senate Commerce Committee Chairman Ted Cruz, and Senator Amy Klobuchar are negotiating a separate bipartisan AI safety bill. One proposal under discussion would give the Commerce Secretary the power to request proof that AI developers are taking reasonable steps to prevent harm — a legal concept known as a “duty of care”.

### The Great American AI Act

Separately, Reps. Obernolte and Trahan released a **269-page discussion draft** of the “Great American AI Act” in June 2026. It would create a comprehensive federal framework for AI governance, assign primary regulatory authority to the Center for AI Standards and Innovation within the Department of Commerce, and impose audit and incident reporting requirements on frontier AI developers.

---

## Why Congress Might Not Give It to Them

So the companies want regulation. The proposals are on the table. Why isn’t Congress acting?

The answer is a combination of presidential opposition, partisan deadlock, a ticking clock, and a fundamental lack of consensus on what regulation should actually look like.

### The Trump Factor

Let’s start with the most obvious obstacle: **President Donald Trump**.

On Monday, September 14, Trump responded to the mounting calls for AI regulation with characteristic bluntness. “AI taking over the World, destroying Humanity, and all other things bad, is a HOAX,” he wrote on social media. He argued that “the only control or ‘guardrails’ that AI needs” is a “strong and smart” president.

Trump’s former AI czar, David Sacks, has favored a light-touch federal framework and warned that a patchwork of state laws could stifle innovation.

Vice President JD Vance has also raised questions about the motives behind the industry’s push for regulation. He questioned why AI companies are asking the government to regulate them, advocating instead for a unified approach between Congress and the White House that maintains America’s competitive edge over China.

The White House has an **opaque system** in place where AI companies can voluntarily submit their latest models for government review up to 30 days before public release. But the eligibility requirements and the review process have not been publicized. It’s soft regulation — executive signals rather than enforceable rules.

### The Partisan Divide

Even if Trump weren’t opposed, Congress is deeply divided.

Democrats want more robust regulation. On Tuesday, House Democrats convened in private to discuss AI. Senate Bernie Sanders is hosting colleagues for a briefing with experts. Top candidates on the ballot this year, like Sen. Jon Ossoff of Georgia, are pressing for sweeping action. Former President Barack Obama is encouraging those with White House aspirations to develop a “very clear plan” to address AI concerns.

But Republicans are skeptical. House Speaker Mike Johnson has largely deferred to the White House. He said on Sunday that AI developers, not Congress, bear primary responsibility for ensuring their products are safe.

“There is an obvious corporate responsibility that the people who are creating these models have to have to ensure that their products are safe,” Johnson said.

He also warned against rushing: “If Congress just races in and does some sort of emergency session to try to regulate AI, we will lose the race to China”.

But Johnson has also shown signs of openness to regulation that clash with Trump’s position, signaling that the Speaker may be willing to explore guardrails even as the President calls them a hoax.

### The Clock

Here’s the practical problem: **the House is set to enter recess at the end of the week until after the midterm elections in early November**.

Some leading Democrats are calling for that recess to be delayed until AI safeguards can be passed. House Minority Leader Hakeem Jeffries said Congress “should not leave town until something is done decisively to protect the safety and the wellbeing of the American people”.

A group of House Democrats wrote a letter to Speaker Johnson calling for more time. “Reasonable minds may disagree about precisely how Congress should regulate this rapidly evolving technology,” they wrote. “We cannot disagree about the imperative for Congress to act”.

But it’s highly unlikely the recess will be delayed. Johnson gave no sign he intended to move quickly. When asked if the House would take action before breaking for the midterms, he replied: “This is a very complex issue”.

### No Consensus

Even among Democrats, there’s no unified position. And there are numerous competing proposals on Capitol Hill: independent reviews, a federal “kill switch,” and a total moratorium on data centers proposed by Senator Bernie Sanders.

These diverging proposals point to a lack of consensus — but also an urgency among some lawmakers to take action.

Senate Majority Leader Thune is locked in negotiations with Klobuchar over a bipartisan AI safety bill. But when asked if an AI bill could pass by the end of the year, he said: “It’s probably TBD”.

### The State Preemption Fight

There’s another layer of complexity: **federal preemption**.

The White House released a National Policy Framework for AI in March 2026 that proposed Congress adopt legislation broadly preempting state AI laws deemed to impose “undue burdens”. Trump also issued an executive order directing the Attorney General to establish an AI Litigation Task Force to challenge state AI laws.

But the industry “lost loudly” an effort to get Congress to ban state-level AI regulation, according to Amba Kak of the AI Now Institute.

Democrats are pushing back. Senators Chris Murphy and Brian Schatz introduced the GUARDRAILS Act to reverse Trump’s state AI law moratorium.

This fight over preemption is a major sticking point. Even if Congress could agree on safety standards, there’s no agreement on whether those standards should override state laws.

---

## The Lobbying Offensive

While the AI companies are publicly calling for regulation, they’re also spending record sums to shape what that regulation looks like.

OpenAI nearly **doubled its federal lobbying expenditure** to a record **$2.22 million** in the first half of 2026. Anthropic nearly **tripled its spending** to **$3.53 million**. Google and Microsoft spent as much as in any quarter since 2020.

The top issues on the lobbying agenda include:

- Federal rules governing the release of new models

- Policy around the construction and power for massive data centers

- Rules governing whether developers make the underlying parameters for their models publicly available

“The lobbying offensive has been as much about deterring regulation as making the case for an affirmative government industrial policy that supports the industry,” said Amba Kak of the AI Now Institute.

Joseph Hoefer of Monument Advocacy put it well: “It’s an export-controlled technology like semiconductors, an infrastructure build-out like telecoms or power, a procurement relationship like defence, and a liability question like the social media platforms, all at the same time. No single playbook covers that”.

The industry “lost loudly” on the state preemption fight, Kak said. “But beyond that, very little else has moved in terms of regulation. Overall, I would say it’s been a good quarter for them”.

---

## The “Regulatory Capture” Warning

Not everyone is convinced that the AI companies’ embrace of regulation is genuine.

Critics — including venture capitalists, academics, and Trump’s former AI czar — have expressed suspicion of **financial and personnel ties** between some monitoring organizations and the big AI labs. They allege that the proposed safety scheme would mainly serve to **cement OpenAI and its top rivals atop the industry**.

The concern is “regulatory capture” — the idea that the companies being regulated end up controlling the regulators. If the AI labs are the ones proposing the rules and funding the bodies that enforce them, can those bodies really be independent?

Zack Kass, a former OpenAI executive, put it more bluntly. He said major AI firms “want to pass cartel-like policy” to protect their own interests.

Chamath Palihapitiya, the venture capitalist, made a similar point about Amodei’s essay. “Dario makes the case to stop open source and concentrate enormous technological and economic power with Anthropic,” he wrote on X.

The suspicion is that the safety argument is being used as a competitive weapon — a way to raise barriers to entry and lock in the dominance of the biggest players.

---

## What the Polls Say

The public is not on the side of inaction.

An NBC News poll found that around **70% of voters disapprove** of President Trump’s handling of the Iran war. That’s a different issue, but it reflects a broader anxiety about the direction of the country and the competence of its leadership.

More directly, a recent poll found that the American people **overwhelmingly want to ban artificial superintelligence** and **pause the development of AI until clear safety standards are established**.

The political pressure is building. But Congress is not responding.

---

## The Human Element: The People Behind the Warnings

Let me bring this back to the people who started this whole conversation.

On September 8, 2026, a 27-year-old researcher named **Jacob Coxon** resigned from Anthropic. He had worked on pretraining at both OpenAI and Anthropic. He quit two months before his equity was due to vest. He walked away from what could have been a life-changing payday.

Here’s what he wrote: **“The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt.”**

His post went viral. And then his former colleagues started speaking up.

**Evan Hubinger**, Anthropic’s alignment science lead — the person whose job is to prevent AI from going rogue — said he estimates there’s a **more than 10% chance** that AI could kill all humans within the next decade.

**Anna Wang**, an Anthropic employee, said she works there because she thinks she can do better at reducing risks from the inside.

**Aidan Clark**, OpenAI’s VP of Research, posted: **“For the first time, I am asking myself if things are moving too fast. I’m honestly not sure.”**

**Bilal Chughtai** resigned from Google DeepMind, saying alignment research isn’t keeping pace with capabilities development.

These are not fringe voices. These are the people who built the technology. And they’re telling us, in public, that they’re scared.

The AI companies heard them. Dario Amodei published his essay. Sam Altman delayed OpenAI’s IPO. The industry started discussing a self-regulatory body.

But Congress? Congress went on recess.

---

## What Comes Next

So where does this leave us?

**If you’re an AI company:** You’ve publicly asked for regulation. Now you need to prove it’s not just a PR move. The third-party evaluators need real power. The safety audits need real teeth. And the standards body needs real independence.

**If you’re a lawmaker:** The clock is ticking. The House is about to leave for the midterms. The Senate is negotiating. The White House is opposed. If you believe AI safety is a defining issue of our era, you need to act before the window closes.

**If you’re an investor:** The regulatory uncertainty is a risk factor. The FRONTIER Act, the Great American AI Act, and the Senate negotiations could all reshape the landscape. Companies that embrace safety standards early may be better positioned than those that fight them.

**If you’re just an American:** The people building AI are telling you they’re scared. They’re telling you there’s a meaningful chance their technology could end human civilization. And they’re telling you they don’t have a plan to prevent it.

That’s not fearmongering. That’s a warning. And right now, Washington isn’t listening.

---

## The Bottom Line

The top AI companies are asking for supervision. They want independent audits. They want embedded watchdogs. They want a standards body. They want regulation.

But Congress might not give it to them.

President Trump is calling AI safety a “hoax.” House Speaker Mike Johnson is deferring to the White House. The House is about to go on recess. And there’s no consensus on what regulation should look like.

The industry says it wants guardrails. The public says it wants guardrails. But the political system is deadlocked.

The question is no longer whether AI needs supervision. The question is whether Washington can actually deliver it — and what happens if it can’t.

As Max Stier of the Partnership for Public Service put it: **“The signals we’re receiving from our political leaders are not representing an understanding of the crisis that I think we are in fact in. Delay is going to prove very costly”**.

---

## Frequently Asked Questions (FAQs)

### 1. What are the top AI companies asking for?

OpenAI, Anthropic, and Google are asking for **embedded third-party evaluators** with employee-level access to their systems, **mandatory safety audits**, a **FINRA-style industry standards body**, and a federal **duty of care** requirement for AI developers.

### 2. What is the FRONTIER Act?

The FRONTIER Act (Frontier Risk Oversight, National Transparency, Independent Evaluation, and Reporting) is a bipartisan House bill sponsored by Reps. Jay Obernolte (R-Calif.) and Lori Trahan (D-Mass.). It would require major AI developers to publish frameworks for assessing catastrophic risks, disclose safety incidents, and submit to independent audits.

### 3. Why is Congress not acting on AI regulation?

Several factors: **President Trump opposes regulation** and has called AI safety a “hoax.” **House Speaker Mike Johnson** says AI companies bear primary responsibility for safety. The **House is about to go on recess** until after the midterms. And there’s **no consensus** among lawmakers on what regulation should look like.

### 4. What is the “regulatory capture” concern?

Critics worry that the AI companies’ embrace of regulation is a form of **regulatory capture** — that the labs are proposing rules and funding the bodies that enforce them, which would cement their dominance and raise barriers to entry for competitors.

### 5. What are the AI companies spending on lobbying?

OpenAI nearly doubled its federal lobbying spending to **$2.22 million** in the first half of 2026. Anthropic nearly tripled its spending to **$3.53 million**. Google and Microsoft also increased their lobbying expenditures.

### 6. What is the Great American AI Act?

The Great American AI Act is a **269-page discussion draft** released by Reps. Obernolte and Trahan in June 2026. It would create a comprehensive federal framework for AI governance, assign regulatory authority to the Center for AI Standards and Innovation, and impose audit and incident reporting requirements.

### 7. Why did Jacob Coxon resign from Anthropic?

Coxon, a 27-year-old researcher, resigned on September 8, 2026, warning that AI could “kill us all by the end of the decade.” He quit two months before his equity was due to vest, walking away from a significant payday. His post went viral and triggered a wave of public statements from other AI researchers.

### 8. What happens if Congress doesn’t act?

If Congress doesn’t act, the U.S. will continue to rely on **soft regulation** — voluntary model submissions, executive signals, and a patchwork of state laws. The AI industry will continue to operate with few enforceable standards, and the risks identified by researchers will remain unaddressed.

### 9. What is the Senate negotiating?

Senate Majority Leader John Thune, Commerce Committee Chairman Ted Cruz, and Senator Amy Klobuchar are negotiating a bipartisan AI safety bill that would include a **duty of care** requirement, giving the Commerce Secretary power to request proof that AI developers are taking reasonable steps to prevent harm.

### 10. What does the public think about AI regulation?

The public overwhelmingly supports regulation. A recent poll found that Americans **want to ban artificial superintelligence** and **pause AI development until clear safety standards are established**. Around 70% of voters disapprove of President Trump’s handling of related issues.

---

## Disclaimer

*This article is for informational and educational purposes only and does not constitute financial, investment, legal, or professional advice. The views expressed are based on publicly available information, including reports from Politico, CNN, the BBC, the Associated Press, Axios, the Financial Times, and other cited sources as of September 16, 2026. Legislative proposals, political dynamics, and regulatory frameworks are subject to rapid change. The author does not endorse any specific policy positions, investment strategies, or companies mentioned. Before making any decisions based on the content of this article, please consult with qualified professionals who can evaluate your specific situation.*