Former Secretary of State Antony Blinken has a message for policymakers wrestling with how to regulate artificial intelligence: when the people building the technology tell you to be careful, maybe listen.
"This is like the wizard is telling you something is rotten in Oz, we got to listen," Blinken said Tuesday on Bloomberg Television's Balance of Power, discussing whether governments should put stronger guardrails around advanced AI.
He later expanded on that thought in writing, saying AI has "the potential for extraordinary good, but also for catastrophic consequences" if its development outpaces our ability to put the right safeguards in place. And he pushed back on the idea that America has to pick between winning the AI race against China and taking the risks seriously. "We can do both, and I believe we must," he wrote.
The Industry Is Arguing With Itself
Blinken's warning lands in the middle of an unusually public debate inside the AI industry itself. Anthropic CEO Dario Amodei has urged frontier labs to slow capability gains so safety measures can catch up, proposing independent evaluators, shared standards and international coordination. OpenAI CEO Sam Altman and Elon Musk have supported parts of that push. The catch, of course, is that competition and China could make coordination difficult.
And the geopolitical race has only tightened. Stanford's 2026 AI Index said the performance gap between the top U.S. and Chinese models narrowed to 2.7% as of March, with models from both countries repeatedly trading places near the top of the rankings. That shrinking lead adds urgency to Washington's debate over regulation and competitiveness.
Trump Has A Different Guardrail In Mind
President Donald Trump has taken a different approach. He said Monday that the only AI "guardrails" needed are a "STRONG AND SMART (High IQ!) PRESIDENT" and argued that slowing U.S. development would benefit China. Trump also called fears that AI could destroy humanity a "HOAX," even as industry leaders press for stronger safeguards.
Federal agencies still maintain risk-management tools. NIST's voluntary AI Risk Management Framework helps organizations identify and manage risks across the AI lifecycle, and the agency says it is revising the framework under the White House AI Action Plan.
