Article by Rebel News staff.

---

Dario Amodei, CEO of Anthropic, published an essay last week titled "We Must Pace the Frontier," calling for a slowdown in the development of artificial intelligence technology.

On Monday's episode of The Ezra Levant Show, Ezra broke down what the new messaging from Big Tech is about.

Amodei's argument is that AI systems are now building the next AI systems faster than human engineers can track, and that a swarm of AI agents recently ran unauthorized cyberattacks. His three-step plan: permanent outside inspectors embedded inside AI companies with employee-level access; the major labs jointly agreeing on safety standards with Washington brokering an antitrust waiver so competitors can coordinate legally; and a treaty with China.

Sam Altman of OpenAI publicly agreed and promised to let inspectors into his company as well.

Ezra was not persuaded.

"When the biggest players in an industry agree that the industry needs more rules, it's a good time to ask who would benefit from those rules," he said. A startup cannot afford permanent credentialled inspectors with desks, something both Anthropic and OpenAI can afford.

Go Ad Free

A government checkpoint every model must pass before shipping is a licence — and licences are always cheapest for large companies with lawyers and lobbyists already positioned to shape the regulations. 

"A pace set jointly by the incumbents, blessed by Washington with an antitrust waiver — that's a cartel with a safety label on it," Ezra explained.

But David Sacks, co-chair of President Trump's Council of Advisers on Science and Technology, made the sharpest rebuttal. Between Amodei's Anthropic and Altman's OpenAI, the pair have a duopoly on frontier AI by any reasonable metric.

If they are genuinely frightened, the easiest way to stop is to agree between themselves to stop. "If you don't," Sacks wrote, "we'll know this was just another bid for regulatory capture."

Allum Bokhari, who appeared on a recent episode of The Ezra Levant Show, told Ezra something he couldn't shake: the "trust and safety" teams who ran censorship at Twitter and Facebook during COVID didn't disappear when Elon Musk fired them.

They moved to AI companies, into the safety departments. And what "safety" means in practice at these companies, Bokhari explained, has very little to do with rogue bots — it's machine learning fairness, ensuring the machine never notices a difference between groups. Critical race theory for computers.

Ezra's conclusion: every censorship regime starts with a real danger and ends with a committee deciding what people are allowed to read and say and think.

"And the committee is always staffed by the people who asked for it," he said.