House Speaker Mike Johnson says the country's top artificial intelligence executives need to sit down with Congress and the White House soon to hammer out safety rules, but he is still refusing to back any pause on AI development. Speaking on CNN, Johnson said the government must move quickly given warnings that advanced AI systems could pose what he called an end-of-humanity threat.
Johnson Draws a Line Between Urgency and a Freeze
According to the New York Post, Johnson said AI leaders, Congress, and the White House need to come together as soon as possible to determine the right balance between AI safety and national security. He argued that U.S. national security must be weighed against the immediate task of making sure AI models are safe. Still, Johnson made clear he will not support a moratorium on artificial intelligence development, a stance he reiterated on CNN Sunday, according to the same report, as reported by RTL Today, arguing that clamping down too hard on AI development risks letting China pull ahead in the technology race.
That resistance comes even as pressure mounts from within his own chamber. In a letter sent September 11, Reps. Sam Liccardo, George Whitesides, Lori Trahan, and Ted Lieu urged Johnson to immediately reconvene the House and cancel an upcoming several-week recess until lawmakers pass bipartisan AI safeguards, according to KION. House Democrats described the risk posed by advanced AI as catastrophic, per that same letter. Those Democrats have pointed to two bills as a starting point.
Kill Switch and Frontier Oversight Bills Await Action
One is the AI Kill Switch Act, introduced in July by Reps. Ted Lieu and Nathaniel Moran, which would require frontier AI developers to build in technical shutdown controls and would let the Department of Homeland Security compel model suspensions during emergencies, according to Rep. Ted Lieu's office. The other is the FRONTIER Act, introduced the same month by Reps. Jay Obernolte and Lori Trahan, which would set up a national risk-based framework requiring independent audits and incident reporting for advanced AI models, per Rep. Jay Obernolte's office. A July poll by the AI Policy Institute found 86% of U.S. voters across party lines support federal requirements mandating guaranteed shutdown capabilities for advanced AI models, according to Lieu's office.
The push for congressional action follows a striking moment of agreement among rival tech executives. On Saturday, Anthropic CEO Dario Amodei published a 3,800-word essay calling for a global slowdown in AI development, and OpenAI CEO Sam Altman and xAI owner Elon Musk agreed with his assessment. Altman had earlier been part of the broader AI policy discussion. Amodei had also previously addressed the risks of advanced AI.
Anthropic Flags State-Linked Misuse of Its Own Models
Adding weight to the safety warnings, Anthropic released a threat intelligence report on September 10 revealing it had disrupted operations between December 2025 and August 2026 in which state-linked actors from Russia, China, and Iran attempted to use its Claude models for biological weapons research, automated cyber espionage, and naval targeting, according to Anadolu Agency. The warnings have also drawn attention from academic circles.
White House and States Pull in Different Directions
States are moving ahead on their own, according to Xinhua.
Lawmakers have introduced federal legislation, while states continue to take their own approaches to AI policy. Hoodline previously reported on Sam Altman's earlier D.C. meetings with Johnson and House Democratic leaders over federal AI oversight, part of a monthslong back-and-forth between the industry and Congress over federal AI oversight.
