Speaker Johnson Calls for White House Meeting with AI CEOs to Set Safety Guardrails

House Speaker Mike Johnson said on September 13, 2026, that Congress must resist imposing an emergency moratorium on artificial intelligence, and he proposed summoning the leaders of the major AI platforms to a White House meeting with lawmakers to settle on safety guardrails. Johnson, a Louisiana Republican, made the remarks in an interview on CNN’s State of the Union that aired at 9:00 a.m. ET.

The interview opened the program, which began with a clip of Anthropic CEO Dario Amodei from what CNN described as a cable news exclusive. Amodei told governments they have the power to regulate the technology and to vote on how it is used and developed, and he said industry caution in building AI cannot remain voluntary, because government and elected representatives need to play a role as well. Host Jake Tapper said OpenAI’s Sam Altman and xAI’s Elon Musk have now said they agree with Amodei that development must slow down, and Tapper asked Johnson whether AI poses an existential threat and whether he would heed the calls for guardrails.

Johnson Rejects Emergency Moratorium

Johnson said he is heeding the executives’ request while also calling on them to act responsibly, describing an obvious corporate responsibility on the people creating the models to ensure their products are safe. He commended a post on X from the evening of September 12, 2026, by White House AI czar David Sacks, which Johnson characterized as explaining that the executives now calling for guardrails are the ones pushing the frontier and that they need to take personal responsibility first. At the same time, he said, lawmakers must resist Congress jumping in and imposing an emergency moratorium, and whatever is done has to be done correctly, safely, and wisely.

Johnson said the United States cannot accept a moratorium because China would pull ahead, framing the problem as national security balanced with the immediate security of making sure the models are safe. If Congress raced into an emergency session to regulate AI, he said, the country would lose the race to China, which he called a threat to every American. He said the public should not panic, that the country should handle the technology as it has handled others in the past, and that safety work must not smother American innovation, because the country has to do both things at the same time.

He pointed to his own record on the issue, saying that almost immediately after he received the gavel almost three years ago he assembled a bipartisan House AI task force with an equal number of members from both parties, which deliberated and researched how to put guardrails in place.

White House Meeting Proposal

Johnson said the leaders and providers of the AI platforms, meaning the executives Tapper had named and a few others, a group he put at about seven or eight people in charge of the field, should come together in a meeting with lawmakers to figure out the right balance on guardrails. Asked when such a meeting should take place, he answered: “I’d do it tomorrow.” He said he has spoken with President Donald Trump about the idea, that the group probably should be summoned together at the White House, and that the participants should gather in one room, close the door, and sort the matter out, adding that he will personally press for the meeting.

Johnson said he has met individually and regularly with the AI leaders since becoming speaker, but that each offers a different prescription for what regulation should look like, that they are all competitors, and that each wants to maintain an edge over the others. Congress is obviously less qualified than the people pushing the frontier to know the technology’s ins and outs, he said, so any action has to be a partnership with the industry, the corporations building the models, and policymakers. He said he tells the executives they must do their part and be responsible on their own side.

Kill Switch, Federal Regulator, and Pre-Release Approval

Asked whether he could support a federal kill switch, the creation of a federal AI regulatory agency, or mandatory government approval of frontier AI models before release, Johnson answered: “Maybe all of it. And those are great ideas.” He compared the moment to the rise of nuclear technology, where oversight boards and international partnerships took decades to assemble, and said the country does not have that much time here. Congress is a deliberative body that needs to catch up, he said, while acknowledging a sense of urgency and saying he is committed to acting with that urgency.

Responses From Gallego, Hassett, Trump, and Dean

Sen. Ruben Gallego, an Arizona Democrat, said on the same program that he is calling on Senate leadership to establish a select committee on AI as soon as possible. He said Johnson and Trump are not meeting the moment, argued that AI differs from earlier technologies because it can make decisions on its own, and compared the situation to Dr. Frankenstein asking for help to stop a monster that is loose. Asked whether he would support the temporary ban on AI development that Sen. Bernie Sanders has advocated, Gallego said everything has to be under consideration.

In a clip the program aired of Trump speaking to reporters in Ireland, the president said the United States could put guardrails in place but that a lot of negative voices are raising concerns that should not be raised, and he said that whoever wins with AI wins. Rep. Madeleine Dean, a Pennsylvania Democrat, said on the program’s panel that Johnson brought the House into session during the week of September 13, 2026, while canceling other session weeks, which she called a mistake, and she noted pending AI legislation from Reps. Ted Lieu and Nathaniel Moran.

National Economic Council Director Kevin Hassett described Amodei’s newly published letter as carefully thought through and said it proposes independent observers wired into the companies the way supervisors oversee banks. He said the technology has so far advanced rapidly and safely under the companies’ own guidance and the White House’s cyber leadership, and that the larger risk is a malevolent actor obtaining a more capable AI and using it to create bioweapons. Hassett said he expects meetings during the week of September 13, 2026, with cyber leaders Sean Cairncross and Michael Kratsios to consider next steps, and that Trump wants to hear from members of Congress who hold strong views on the issue.