OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei this week agreed to put third-party safety monitors in their companies. Amodei said in an essay titled 'We must pace the frontier' that 'the stakes are too high for pacing to be an empty exercise — we need to use the time it gives us wisely.'
Altman later backed that idea in an X post, and Elon Musk and Google DeepMind founder Demis Hassabis also voiced support. The debate sharpened after former Anthropic engineer Jacob Coxon warned of a looming AI 'armageddon,' with some leaders saying the industry needs a slowdown to reduce catastrophic risk.
Not everyone is buying the alarm. White House adviser David Sacks said on X, 'go ahead,' but added that companies should 'stop pretending you need anyone else's permission.' He said a slowdown could be a bid for 'regulatory capture' or 'an election-season psyop.'
Hugging Face CEO Clément Delangue said AI safety cannot be left solely to frontier labs. He said his company was launching an 'open alignment initiative' and asked that it be included in Amodei's proposed third-party evaluator program. Venture capitalist Gavin Baker argued that AI can be 'democratize[d] and distribute[d] broadly and safely without centralizing it in the hands of a few corporations.'
The political response is splitting along a familiar fault line: safety versus speed. Senate Democrats have already pushed legislation to ban the development and deployment of artificial superintelligence and temporarily pause advanced AI systems. Senator Sanders said after Coxon resigned from Anthropic that 'Mr. Coxon is right,' and later called on Trump to negotiate a treaty with President Xi of China for a pause on advanced AI development and a ban on super intelligence.
House Speaker Mike Johnson rejected an emergency moratorium on CNN's 'State of the Union,' warning that 'we cannot put a moratorium on this because China will overlap us.' He said Congress should not jump in with an emergency moratorium and instead have the companies 'come together in a meeting with us and figure out what the right balance is.'
Amy Kremer, chairwoman of Humans First, said policymakers should pursue safeguards without surrendering America's technological advantage to Beijing. She called for protecting children, workers and national security, while keeping local control over data-center projects.
Here's the thing: the same companies asking for guardrails are also the ones shipping the frontier models. That makes the safety push look less like panic and more like an attempt to set the rules before regulators do it for them. If the next phase of AI is going to be slower, safer, and more heavily watched, the public should ask who gets to write the playbook — and who pays if the bet goes wrong.
Nobody asked for a race to build the most powerful system first and sort out the consequences later. But that is exactly the logic now driving Washington, Silicon Valley, and the administrative state into the same room.
