Anthropic CEO Dario Amodei publicly urged the AI industry last Saturday to decelerate the pace of frontier model capability improvements, proposing measures such as third-party evaluations, the establishment of industry safety standards, and the coordination of global rules. OpenAI chief Sam Altman and xAI leader Elon Musk subsequently endorsed the direction Amodei outlined. Amodei also stated that Anthropic would adopt new safety protocols, including bringing in external assessors, with Altman indicating he would follow suit.
President Trump's stance, however, differs markedly. During a visit to Ireland on Sunday, he told reporters that the United States must preserve its leading edge in AI. "We're the most advanced country in the world, and frankly, I want to keep it that way because whoever wins AI wins the world," Trump said. While not entirely dismissing safety guardrails, Trump argued the concerns have been blown out of proportion. "We can put guardrails up. We can do this and that, but I think there are a lot of negative forces talking about things that aren't going to happen," he remarked.
Yet the Trump administration has not wholly rejected AI safety measures. Kevin Hassett, Director of the National Economic Council, indicated that Amodei's proposal for enhanced AI safeguards could serve as a template for private sector action. Hassett characterized AI safety as a "solvable problem," noting that while AI development could accelerate beyond existing cybersecurity defenses, companies are also devising new protective measures. David Sacks, Trump's former AI czar who now co-chairs his technology advisory committee, suggested AI companies can slow their own development pace without awaiting government approval. "If your product facilitates a genuinely devastating cyberattack, you'll face enormous product liability risk," Sacks wrote on X. He added that the market is already penalizing models with unpredictable or unauthorized behavior.
The warning that AI "could kill everyone" has propelled safety debates from tech circles onto the political agenda. Previously, Anthropic researcher Jacob Coxon resigned, cautioning that "the people building AI genuinely believe it could kill us all before the end of this decade." Amodei subsequently responded in a public essay, arguing the industry "must slow the pace of improving AI model capabilities." Speaking on CBS's "Sunday Morning," Amodei affirmed that government should have a role in AI regulation. "Regulation gives the public and their elected representatives a voice and limits what private companies can do," he said.
Democrats are pushing AI safety further into the November midterm elections. Senator Chris Coons argued that Congress currently lacks the expertise needed to address rapidly evolving AI. "That's one of our real challenges: some people in the Senate and House aren't as engaged, as curious, or as up-to-date on this fast-moving technology as we need them to be," Coons said. He advocated for Republicans and Democrats to temporarily set aside pre-election partisan wrangling to establish genuine guardrails before AI "spirals out of control." According to reports, former President Barack Obama also told a private fundraiser last week that Democrats should make AI oversight a central campaign issue.
House Minority Leader Hakeem Jeffries said House Democrats would hold a caucus meeting Tuesday to discuss potential AI actions. Speaking on ABC's "This Week," he asserted Republicans have "abdicated their responsibility" and called on Congress to "act decisively now." House Speaker Mike Johnson, while not denying AI safety risks, opposes sacrificing development momentum for safety. He urged against "panicking." "We can't pause this... This is national security, and it must be balanced with the immediate safety of ensuring models are secure," Johnson said. He also called on tech companies and the government to collaboratively find AI guardrails, noting firms have "a corporate responsibility to ensure product safety." Johnson advocated for direct dialogue between AI industry leaders and government officials, suggesting they "go into a big room, close the door, and get it sorted out."
Democrats have sharply criticized the Republican response and demanded swift regulatory action. Arizona Senator Ruben Gallego accused Johnson and Trump of being "out of step with the times" regarding the threats posed by AI. At a campaign rally in Athens, Georgia, Democratic Senator Jon Ossoff blasted Trump's response and accused tech titans of digging "bunkers" while warning of human extinction. He argued that a capable president would act swiftly to protect and reassure the American people by sending inspectors into frontier labs, strengthening national bioterrorism defenses, demanding congressional legislation, and leading the world toward an AI treaty. "But instead, the president and the speaker of the House stand by, hoping for the best," Ossoff said.
The question of returns on massive AI industry investments has further intertwined safety disputes with capital markets. Data center construction is driving up local resource demands, while concerns over AI potentially raising electricity costs, displacing certain jobs, and whether enormous infrastructure spending will ultimately yield sufficient returns have all become pressing issues for American society and investors alike.