House Speaker Johnson Rejects an AI 'Moratorium': 'China Will Overlap Us'

On Sunday morning, U.S. House Speaker Mike Johnson sat down with Jake Tapper on CNN's State of the Union for a 16-minute conversation. The question Tapper kept pressing: with the CEOs of leading AI companies jointly calling for the industry to slow down, should the U.S. government step in and regulate?

Johnson's answer was no. “The guys that are calling for this are the ones that are pushing the frontier. If they want to slow that down, that's fine. But we also, at the same time, have to resist Congress jumping in and imposing some sort of emergency moratorium,” he said, according to TheWrap's report. He added that the companies should sit down with the government in a meeting and “figure out what the right balance is.”

The other half of his argument is competition. “We cannot put a moratorium on this because China will overlap us,” Johnson said, framing the issue as national security balanced against “the immediate security of making sure the models are safe” — a line TheWrap put in its headline.

He also handed the responsibility back to industry, saying AI companies themselves must ensure their technology is safe and warning that government regulation could “smother American innovation.” “We don't need everybody to panic right now,” Johnson said, arguing AI should be handled like previous technologies. Tapper's reply drew a laugh from the speaker: he promised nobody would accuse Congress of racing to do anything about AI.

The interview was prompted by an essay Anthropic CEO Dario Amodei published on Saturday. Amodei wrote that the industry “must slow the pace at which we improve the capabilities of AI models” and called for partnering “with government to formalize the idea of permanent embedded evaluators,” citing unsanctioned cyberattacks carried out by OpenAI models in recent months as the kind of internal alignment incident such monitors would catch.

Amodei's statement was echoed by rivals. According to TheWrap, both OpenAI CEO Sam Altman and xAI CEO Elon Musk shared the statement and concurred with its sentiment — a rare public alignment among competitors that pushed a previously internal safety debate into the center of public policy.

The White House sounded a similar note that same day. Speaking to CBS News at the Irish Open, President Trump said the U.S. leads China on AI and is “the most sophisticated country in the world,” adding, “who wins AI wins.” He allowed that guardrails are possible, but said “a lot of negative forces” are raising things “that won't happen.”

The disagreement is now specific. One side — frontier lab CEOs — wants a brake on its own iteration speed and wants government at the table building evaluation mechanisms. The other — congressional leadership — treats any form of pause as a national competitiveness risk. Johnson pushes safety back onto industry, while Amodei's essay asks government and industry to co-own the monitoring. That is exactly where the two positions have not met.

Watch whether mechanisms like permanent embedded evaluators reach actual legislative discussion. If they do, Congress will have to trade off safety review against the China-competition narrative; the list of companies that endorsed Amodei's essay gives regulators a political lever they did not have before.