When Ronald Reagan sat across from Mikhail Gorbachev in Reykjavik in 1986, the stakes could not have been higher. The future of nuclear arms control – and with it, the safety of the American people – hung in the balance. Reagan negotiated from a position of strength, refused to blink, and ultimately helped bring the Cold War to a close on America’s terms.
As President Trump prepares to travel to Beijing, he faces a moment with similar historical weight. The technology in question is not nuclear warheads. It is artificial intelligence – and the risks are just as real.
Advanced AI models are no longer a distant concern for science fiction writers and Silicon Valley futurists. They are here, they are powerful, and in the wrong hands, they pose a genuine threat to American national security.
Models like Mythos represent a new category of capability – systems that can reason, plan, and act in ways that outpace our current frameworks for oversight and control. The question is not whether AI will reshape the global balance of power. It already is.
The question is whether America will lead that transformation or cede ground to adversaries who have no interest in keeping us safe.
That is why it is genuinely encouraging to see the Trump administration taking this threat seriously.
Treasury Secretary Scott Bessent deserves significant credit for elevating AI safety as a national security priority and for mobilizing key administration officials to act. He understood before many of his peers that this is not just a technology policy question – it is an economic security question, a national defense question, and an American sovereignty question all rolled into one.
I also want to recognize President Trump, White House Chief of Staff Susie Wiles, and National Economic Council Director Kevin Hassett for moving swiftly and decisively. Patriotic leaders inside this administration have been pushing for action, and they are getting results.
Now comes the harder part.
China is not a passive actor. Beijing has made artificial intelligence a cornerstone of its national strategy, pumping billions into frontier model development and making no secret of its ambitions to surpass the United States. The Chinese Communist Party does not share our values around transparency, human rights, or the rule of law – and there is no reason to assume they will self-regulate advanced AI in ways that protect Americans or our allies.
That is precisely why President Trump’s trip to Beijing represents a once-in-a-generation opportunity. Just as Reagan used his credibility and leverage to negotiate arms control from a position of strength, President Trump can use America’s current AI advantages – our talent, our capital, our infrastructure – as leverage to secure a meaningful agreement on advanced AI safety and security.
Not an agreement that trusts China on its word, but one with real verification mechanisms, clear red lines, and consequences for violations.
An “America First” AI deal would accomplish several things at once. It would reduce the risk of an AI arms race that neither side can fully control. It would establish the United States as the global standard-setter on responsible AI development. And it would protect the American people from the catastrophic risks that unchecked advanced AI could pose – whether through accident, misuse, or deliberate adversarial deployment.
President Trump has shown he knows how to walk into a room with an adversary and walk out with a deal that puts America first. He did it on trade. He did it on NATO burden-sharing. Now, he can do it on AI.
The moment is now. The stakes are enormous. And the American people are counting on him to get it right.
Brendan Steinhauser is CEO of The Alliance for Secure AI, a nonprofit organization that educates policymakers and the public about the implications of advanced AI.