Amodei Says China Is Toughest Dilemma for AI Slowdown as Washington Debates Guardrails

Amodei told CBS that the more long-term idea would be working together to put a speed limit on the rate of AI progress. 'I think that's going to be very difficult because the incentives to pull ahead and the military advantage that you get from that are so large,' he said. 'And honestly, I don't know if it's possible, but we should try.'

Anthropic has 'unilaterally' committed to the first step, Amodei said: granting third-party evaluators employee-level access to verify safety practices and report incidents. The second step encourages leading AI companies in democratic countries to coordinate common safety standards. The third calls for coordination between democratic governments and authoritarian governments.

The essay followed warnings from industry researchers about the technology's potential to cause catastrophic harm. Anthropic researcher Jacob Coxon announced he quit, saying the company and rival OpenAI are 'gambling with our lives.' The debate is unfolding as the AI industry faces public backlash over data centers and mounting calls for regulation from Washington. Anthropic and OpenAI are also preparing for potentially historic IPOs, though neither has set an official date.

OpenAI CEO Sam Altman said on Saturday that OpenAI will likely not go public this year, in part because of safety concerns. 'Right now would be an ill-advised moment to go public,' Altman told Fortune. He also expressed support for Amodei's essay on X, writing that he agrees on the need to 'pace the frontier.' TechCrunch reported that OpenAI would commit to 'having independent evaluators with employee-like access.' Google DeepMind chair Demis Hassabis wrote on X that Amodei's essay 'points towards the right path forward,' adding that details need work but the direction is correct. Musk wrote early Sunday that 'Dario is right that there should be some oversight' and that 'peer review of AI by competitors is the right way to start this off.' Musk had been a major critic of Anthropic until SpaceX, which acquired his AI startup xAI earlier this year, signed a major compute deal with Anthropic in May. Anthropic will pay SpaceX $1.25 billion per month through May 2029, according to filings.

Other industry figures weighed in. Alexandr Wang, Meta's AI chief, did not address Amodei's essay directly but shared a post saying Meta Superintelligence Labs is 'rapidly scaling up the share of our efforts that goes into alignment as our models become more powerful.' He said alignment can be the gating factor for scaling near the frontier. David Sacks, a former Trump AI and crypto czar and a critic of Anthropic, wrote that if OpenAI and Anthropic believe their models are 'scary enough' to warrant a slowdown, he supports their 'decision to be responsible.' He also said both companies should 'stop pretending you need anyone else's permission' and acknowledge their motivations are not purely altruistic. Sacks said they face 'massive product-liability exposure' if their products enable something like a damaging cyberattack, making it good business to trade some raw power for reliability and predictability.

Washington is scrambling to respond. Senator Ruben Gallego, Democrat of Arizona, said on CNN's 'State of the Union' that 'Dr. Frankenstein is telling us the monster is escaping; help us stop this,' and that the method Congress is answering with is not meeting the moment. Pennsylvania Governor Josh Shapiro wrote on X that the industry leaders' commentary should be heeded as 'a bright flashing red light telling the President and Congress they need to act now to regulate and rein in AI development.' CNBC reported that the events have left Washington, which has done little to regulate AI, with its back against the wall as midterm elections approach and Congress prepares to leave for campaigning.

On Friday, a group of Democrats led by Representative Sam Liccardo of California wrote to House Speaker Mike Johnson demanding that the House stay in session until it passes meaningful safeguards. 'The House should return to Washington immediately and remain in session until Congress advances meaningful, bipartisan AI safeguards,' the letter said. The Democrats urged consideration of proposals including transparency and evaluation of frontier models, 'kill switch' requirements, and a waiver of antitrust laws so the industry can work together on safety and security. Johnson is scheduled to send the House home after this week until November's elections. House Democratic Leader Hakeem Jeffries said Sunday on ABC's 'This Week' that Democrats would meet on Tuesday to discuss AI guardrails and that they should take decisive action to slow development. Johnson told CNN that an emergency session to regulate AI would mean losing the race to China, calling that a threat to every American. He said Congress should not panic, should handle the technology responsibly without smothering innovation, and called for a meeting between Washington leaders and AI industry leaders. Gallego also called for a summit and said Johnson, President Donald Trump, and Congress are not meeting the moment. He said AI 'does not care once it becomes fully dangerous whether we're Chinese or American.'

Former President Barack Obama urged Democrats to prioritize a public conversation about AI management and safety at a closed-door Manhattan fundraiser, remarks first reported by The New York Times. Obama was interviewed by Jeffries and said that once Democrats regain the House majority, they should 'put together a framework for a very public conversation.' He said AI 'is moving very fast in private hands, and if we don't get on top of it, I think can be dangerous,' while adding that if handled well it could accelerate drug development and help find pathways to clean energy. He also said Democrats should present plans for safety and for making sure children are not corrupted by the technology. TechCrunch reported that Obama called AI one of Democrats' 'central agendas' and said they need 'a very clear plan.' The Guardian described the fundraiser as last week, while TechCrunch said it took place Thursday. Jeffries said Obama 'is correct that decisive action must be taken on artificial intelligence.' The New York Times reported that Obama has served as a sounding board to AI executives and has spoken with Amodei and Altman.

Trump, speaking to reporters Sunday at his Irish golf course, said many 'very negative forces' are raising AI concerns that 'won't happen,' and that his main concern is that the United States remain the industry leader. 'We're leading China in AI. We're the most sophisticated country in the world, and frankly I want to keep it that way because whoever wins AI wins,' Trump said, according to Reuters. He added that guardrails could be put in place but repeated his criticism of negative forces. Trump previously wrote on TruthSocial that communities that do not want data centers want to end up 'backwards and poor,' and said, 'Let Data Reign.' His administration earlier released a legislative framework for AI that would preempt state laws and shift the child safety burden to parents.

Separately, SiliconANGLE reported that Q-Day, the point at which quantum computers begin to break the encryption protecting most of the internet, is approaching with most organizations unprepared. Google set 2029 as its own migration deadline for post-quantum cryptography, ahead of the National Security Agency's 2031 target and the National Institute of Standards and Technology's 2035 guideline for national security systems. Google said it accelerated its timeframe because quantum hardware and error correction are advancing faster than expected. A March paper by researchers from Caltech, UC Berkeley, and quantum startup Oratomic found that a fault-tolerant quantum computer running Shor's algorithm may require as few as 10,000 to 26,000 qubits, roughly two orders of magnitude below the millions once assumed. Public-key cryptography underpins TLS handshakes, digital signatures, bank key exchanges, and server certificates, meaning the layers break for everyone at once. The report said piecemeal patching across thousands of applications is not realistic, and argued for post-quantum readiness at the network layer. DigiCert's 2026 Quantum Readiness Outlook found that 87% of organizations are planning, testing, or implementing post-quantum initiatives, but only 7% have deployed quantum-safe or hybrid cryptography across a meaningful share of their digital certificates. Axiad IDS found that 51% of enterprise security leaders have never formally tested their public-facing infrastructure for post-quantum key exchange, and nearly half lack a named leader responsible for the effort.