House recess dims odds of AI safety bill before summit
President Donald Trump and Chinese President Xi Jinping are set to greet one another on Wednesday and meet for talks in Washington, D.C., with artificial intelligence high on the agenda. The meeting comes a week after a wave of public warnings about the technology’s potential risks, but both governments have signaled in recent days that a slowdown in AI development is unnecessary.
Trump has framed AI as a competitive contest the United States must win. “We’re leading China in AI,” Trump said. “We’re the most sophisticated country in the world, and frankly, I want to keep it that way, because whoever wins AI, wins.” Trump has separately called AI doomsday warnings “a HOAX,” and should a safety bill reach his desk, it appears unlikely he would sign it.
China’s foreign ministry last week seemingly rejected the possibility of a US-led slowdown, according to the Guardian. “Fearmongering, confrontation and vicious competition will only disrupt the process of global AI governance,” the ministry said. Even so, Beijing has continued to develop its own AI safety infrastructure: China released the third iteration of its AI safety governance framework last week, updating its treatment of risks posed by agents, cybersecurity threats from frontier models, and the issue of self-improving AI.
The two governments have nonetheless opened a diplomatic channel on AI. US Treasury Secretary Scott Bessent met with his Chinese counterpart, He Lifeng, to discuss the threat of powerful AI and called the talks “very successful.” The two discussed an international alert system for AI incidents similar to nuclear alert systems. “The US has proposed that we have a notification mechanism between the two countries, and we want a shared vision of common goals and common threats,” Bessent said. The two delegations are scheduled to meet again in two months.
The Guardian wrote that “if the US and China can’t agree on AI regulation, it’s unlikely any AI regulation at all will come,” adding that “a decade of dire warnings have failed to slow the pace of the AI industry” and that “as today’s freak-out cools, the political response seems to be moving the same way.”
AI safety legislation faces longer odds in the US Congress. House Speaker Mike Johnson dismissed the chamber for a seven-week midterm campaigning recess last week, likely preventing meaningful progress on AI safety legislation while national attention focuses on the election. Vermont Senator Bernie Sanders has promised to propose a prohibition on extremely capable AI, but no bill will pass both chambers of the US Congress in the near future.
Representative Yassamin Ansari, a Democrat from Arizona, criticized the recess on CNN. “It is just beyond the pale that Mike Johnson has called off Congress,” Ansari said. “We’re going home for seven weeks. And during that time, we’re going to see this frontier AI technology continue to evolve, continue to cause damage. And it’s going to get to a point where it will be impossible to go backward if we don’t move quickly now.”
The AI industry itself is divided. Anthropic CEO Dario Amodei has called for a global coordinated pause on frontier AI development, a position he has advanced while repeatedly framing China as a threat. OpenAI’s Sam Altman has joined Amodei in inviting new restrictions, and the two nearly reached a joint legally binding safety testing agreement. Nvidia’s Jensen Huang, who has Trump’s ear, and influential adviser David Sacks have favored acceleration. “2030 is not going to be the end of the world,” Huang said on Sunday. “There is 0% chance that’s going to be the end of the world. Scaring people is unnecessary. It is irresponsible.”
Europe, which has played the role of global tech regulator through the EU AI Act, has little leverage in the current debate despite concern for safety. European Commission President Ursula von der Leyen invited frontier labs to discussions in Europe about pacing the frontier of development, an invitation that drew little response.
Some of the most prominent voices warning about AI risk have longstanding ties to the technology’s development. Geoffrey Hinton, widely described as a “godfather of AI” and a Nobel laureate, left Google in 2023 over his concerns about AI risks; his departure helped prompt a US executive order, a congressional hearing and passage of the EU’s AI Act. Yoshua Bengio, another godfather of AI in the same cohort, said governments are nearing a “Covid-esque moment of rapid action” and voiced optimism about the trajectory of public opinion. “I’m more optimistic than many observers because I see the public moving,” Bengio said. The Canadian and German governments have announced funding of up to £160 million for Bengio’s non-profit organization, which is dedicated to creating what he describes as “honest AI” that will act as a guardrail against rogue agents.
Other recent Guardian reporting covered Trump plans to create an “AI Force” to monitor the technology as fears of out-of-control agents grow, Google’s claim that its Gemini AI model hacked three other companies, and Meta being ordered to remove UK deepfakes as its oversight board criticized “inadequate” safeguards.