OpenAI Says It’s Working With Anthropic and Google on AI Safety — And That Changes Everything

 

OpenAI Says It’s Working With Anthropic and Google on AI Safety — And That Changes Everything

**In a stunning move that would have been unthinkable a year ago, OpenAI has confirmed it’s working with its two biggest rivals—Anthropic and Google DeepMind—to build a shared safety framework for artificial intelligence. The goal is nothing less than preventing a global catastrophe. But can three companies that compete for everything actually cooperate on the one thing that matters most? And why is the industry moving faster than the government?**

---

## The Announcement That Shook the AI World

Let me tell you what happened on Tuesday, September 15, 2026, because it’s the kind of moment that makes you sit up and pay attention.

**Chris Lehane**, OpenAI’s global policy chief, was in Washington. He was there to talk to lawmakers about the existential risks posed by the technology his company is building. And during a briefing, he confirmed something that had been rumored for weeks: **OpenAI has been working with Anthropic and Google DeepMind on AI safety.**

Not competition. Not beating them to the next model. Working with them.

“It’s better to try to work together to prioritize safety,” Lehane said.

This isn’t just a press release. This is a fundamental shift in how the AI industry operates. And it’s happening because the people building the technology are scared.

---

## The Catalyst: One Researcher’s Resignation

To understand why this is happening now, you have to go back a week.

**Jacob Coxon**, a 27-year-old researcher, quit his job at Anthropic. He had worked on pretraining at both OpenAI and Anthropic. He resigned two months before his equity was due to vest. He walked away from a life-changing payday.

And then he posted on X: **“The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt.”**

The post went viral—more than 170 million views. And then something remarkable happened. His former colleagues agreed with him.

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

**Aidan Clark**, OpenAI’s VP of Research, posted: **“For the first time, I am asking myself if things are moving too fast.”**

**Bilal Chughtai** resigned from Google DeepMind, warning: **“I earnestly believe that AI has the potential to kill us all.”**

The message was impossible to ignore. The people who know the most about what’s being built are the most afraid of it. And they’re done staying silent.

---

## The Second Catalyst: Dario Amodei’s Essay

Then, on Saturday, **Dario Amodei**, Anthropic’s CEO, published a **3,800-word essay** calling for the AI industry to deliberately slow down. He warned that AI could be capable within six to twelve months of leading a swarm that could take over the entire internet. He proposed independent third-party evaluators embedded in AI companies, antitrust waivers for coordination on safety standards, and international cooperation.

The response was immediate and surprising.

**Sam Altman** agreed. **Elon Musk** agreed. **Demis Hassabis** of Google DeepMind agreed.

For the first time, the leaders of the most powerful AI companies in the world were publicly saying the same thing: we need to pump the brakes.

But then the other shoe dropped.

**Jensen Huang**, the CEO of Nvidia, pushed back. At the All-In Summit in Los Angeles, he said: **“The idea that AI could destroy the world has no scientific basis.”**

Then his phone rang. It was President Trump. And Trump delivered a message that echoed Huang’s own philosophy: **“AI taking over the World, destroying Humanity, and all other things bad, is a HOAX.”**

The battle lines were drawn. And then OpenAI did something that changed the game.

---

## The Collaboration: What’s Actually Happening

So what does “working together on AI safety” actually mean?

According to multiple reports, the three companies have been in discussions for **several weeks** about creating a shared safety framework. The centerpiece of those discussions is an **industry standards body**—a self-regulating organization modeled on the **Financial Industry Regulatory Authority (FINRA)**, which oversees U.S. brokers and investment firms.

The idea originated with **Demis Hassabis**, Google DeepMind’s CEO. In July 2026, he published a proposal calling for a U.S.-led standards body that would test the most powerful AI systems before they are released to the public.

Here’s how he envisioned it:

- **Public-private partnership**: Overseen by the government but funded by the industry

- **Independent experts**: Staffed by “independent leading technical experts and open-source representatives”

- **Substantial funding**: “Funding would need to be substantial and likely mostly come from industry, in order to attract world-class technical talent and provide the necessary compute resources for large-scale testing”

The conversations are ongoing, and they’re continuing **with or without the Trump administration**.

---

## The Big Question: Do They Need Antitrust Waivers?

Here’s the thing about three competitors coordinating on anything: it raises antitrust questions.

**Dario Amodei** had requested limited antitrust exemptions from the government to allow AI companies to coordinate specifically on safety risks. But Lehane said OpenAI **does not believe** the three companies need an antitrust waiver to coordinate on safety matters.

“It’s better to try to work together to prioritize safety,” he said.

The fact that they’re moving forward without government permission is significant. It shows the industry is willing to act—even if Washington won’t give them cover.

---

## The Skeptics: Is This a Cartel?

Not everyone is convinced this collaboration is about safety.

**Aidan Gomez**, the CEO of Canadian AI company Cohere, accused the dominant American AI labs of forming a **“cartel”** under the guise of safety.

“AI needs guardrails. That is not the dispute and never has been,” Gomez wrote in a blog post on Sunday. “The dispute is over who writes them, who gets to participate and whose interests the rules are protecting”.

That’s the uncomfortable question. If the biggest AI companies are the ones writing the safety rules, will those rules protect the public—or protect their own dominance?

**FTC Chairman Andrew Ferguson** said Tuesday that everyone should be **“deeply suspicious”** of AI companies seeking antitrust exemptions while lobbying for new regulations. He criticized the plan as an attempt to “deepen the moat” and protect existing dominant players.

And **Senator Bernie Sanders** said tech leaders could not be trusted to keep promises to slow development or set safety standards.

“When the future of humanity is at stake, we need binding international safety rules, not voluntary standards from the industry,” he said at an AI safety event in Washington.

---

## The Political Divide: Trump vs. the AI Companies

The collaboration is happening against a backdrop of intense political division.

**President Trump** has made his position clear: AI safety is a “hoax.” He believes regulation would hand an advantage to China. He wants the U.S. to accelerate, not slow down.

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

**Mark Zuckerberg** reportedly advised Trump against the idea of a national AI regulator in a phone call this summer.

But even within the administration, there’s skepticism about the AI companies’ motives. 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.

So the AI companies are moving forward without the government. They’re building their own safety standards. And they’re doing it because they believe the risk is too great to wait.

---

## Why This Matters for American Investors

Let’s bring this down to earth. What does this collaboration mean for your portfolio?

**For Nvidia investors:** Jensen Huang’s pushback against the slowdown camp is strategically important. The biggest risk for Nvidia is regulation that would limit AI chip purchases or data center construction. By aligning with Trump and rejecting the “doom” narrative, Huang is protecting his company’s core business. Barron’s noted that his position is **“probably the best thing he can do for the stock right now.”**

**For OpenAI, Anthropic, and Google investors:** The collaboration signals that the industry is trying to get ahead of regulation. If the standards body succeeds, it could provide a framework that allows AI development to continue while addressing public concerns. If it fails, or if it’s seen as a cartel, it could invite stricter government intervention.

**For the broader market:** The AI trade has been the engine of this bull market. The outcome of this safety debate will shape the regulatory environment for years to come. Companies that embrace safety standards early may be better positioned than those that fight them.

---

## The Bottom Line: A Race Against Time

The collaboration between OpenAI, Anthropic, and Google is a recognition of something profound. The people building the most powerful technology in human history believe it could end human civilization. And they don’t have a plan to prevent it.

They’re trying to build one. They’re trying to do it together. And they’re doing it while the government is distracted, divided, and in some cases, openly hostile to the idea of regulation.

As **Anna Wang**, an Anthropic employee, put it: **“There is not yet a viable scientific plan to solve risks from recursively self-improving AI. Please look up.”**

That’s the warning. And it’s coming from the people building the technology.

The question is whether anyone is listening.

---

## Frequently Asked Questions (FAQs)

### 1. What did OpenAI announce about AI safety?

OpenAI’s global policy chief, Chris Lehane, confirmed on September 15, 2026, that OpenAI has been working with Anthropic and Google DeepMind on AI safety for several weeks. The goal is to create a shared safety framework and potentially an industry standards body.

### 2. Why are these competitors working together?

They’re working together because the people inside these companies—researchers, engineers, and executives—believe there’s a meaningful risk that AI could cause catastrophic harm. A researcher who resigned from Anthropic warned that AI could “kill us all by the end of the decade,” and his former colleagues publicly agreed.

### 3. What is the industry standards body they’re discussing?

It’s a proposal modeled on FINRA (the Financial Industry Regulatory Authority), which oversees U.S. brokers and investment firms. The AI standards body would test advanced AI models before deployment. It would be a public-private partnership, overseen by the government but funded by the industry and staffed by independent technical experts.

### 4. Do they need antitrust waivers to coordinate?

OpenAI’s Chris Lehane said the companies **do not believe** they need an antitrust waiver to coordinate on safety matters. Dario Amodei had requested limited antitrust exemptions, but the companies are moving forward without them.

### 5. Why is President Trump opposed to AI regulation?

Trump has called AI safety a “hoax” and argued that slowing down development would hand an advantage to China. His former AI czar, David Sacks, has favored a light-touch federal framework. Mark Zuckerberg reportedly advised Trump against creating a national AI regulator.

### 6. What are the critics saying?

Critics, including Cohere CEO Aidan Gomez, accuse the dominant AI labs of forming a “cartel” under the guise of safety. FTC Chairman Andrew Ferguson said everyone should be “deeply suspicious” of AI companies seeking antitrust exemptions. Senator Bernie Sanders said voluntary industry standards are not enough.

### 7. What do the AI companies say about the risks?

AI researchers have publicly warned about catastrophic risks. Evan Hubinger, Anthropic’s alignment science lead, estimates a **more than 10% chance** of human extinction from AI within a decade. Anna Wang, another Anthropic employee, says there is **“no viable scientific plan”** to solve the risks.

### 8. What happens next?

The conversations between OpenAI, Anthropic, and Google are ongoing. They’re continuing with or without the Trump administration. OpenAI’s chief scientist, Jakub Pachocki, said they’ve been talking to external organizations about “potential concrete standards” and will have more to share “in the next months.”

### 9. How does this affect investors?

For Nvidia, Jensen Huang’s opposition to slowdown calls protects the company’s core business. For OpenAI, Anthropic, and Google, the collaboration could help them get ahead of regulation. For the broader market, the outcome will shape the AI trade for years to come.

### 10. What can I do with this information?

Stay informed. Pay attention to AI regulation debates. Ask your representatives where they stand. And understand that the people who know the most about this technology are telling us it’s dangerous. The decisions being made now will shape the future for everyone.

---

## 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 Bloomberg, CNN, Reuters, the Associated Press, and other cited sources as of September 16, 2026. The field of AI safety is rapidly evolving, and the risks and probabilities discussed are estimates that may 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.*