AI guardrails can’t wait — vote Republicans out in November

You could easily have missed it, but history books — assuming, of course, that we are still around to write them — may well record Sept. 12, 2026, as a watershed moment in human history.

That’s the day Anthropic CEO Dario Amodei, the company that currently runs the most advanced AI model in the world, published an essay warning of the dangers of runaway artificial intelligence. In it, he called for global regulation and slowing the pace of AI development.

It’s a serious and thoughtful essay that balances the need to develop AI in a competitive environment against the risk that unchecked AI will “escape containment” and do serious, possibly civilization-ending, damage in the real world. Amodei estimates that rogue AI agents are less than a year away from being able to permanently destroy the internet.

And what was the American government’s official response to all this? Here’s what President Trump had to say . “The only control or ‘guardrails’ that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT, and the U.S.A. has that, in spades!”

This is beyond parody. Trump, famously, doesn’t even use a computer . Pretending that he will be single-handedly responsible for regulating AI isn’t a policy, it’s a punchline.

You can’t use “common sense” to regulate AI. Anyone who understands anything about AI understands that no one knows what even the current versions of AI are capable of. Amadeo doesn’t know. Elon Musk doesn’t know . Sam Altman doesn’t know . That’s why the recent incident where an Open AI model attacked a rival company was so disturbing.

AI possesses what are known as “ emergent properties, ” which makes it impossible to fully predict what an AI model will do in any situation. It’s absolutely critical that everyone cooperates in understanding what is going on inside these models — and that we create effective mechanisms to keep AI under human control.

This isn’t just a domestic problem. In any sane world, there would have been international negotiations to regulate AI starting at least a year ago. There is no question that AI poses as much of a threat to humanity as nuclear weapons do.

But perhaps a better analogy would be biological weapons. Every country regulates and restricts what sorts of biological research individuals and private companies can do. There’s even an international convention banning biological weapons. Just like with dangerous pathogens, you may be researching AI in what you believe is a controlled environment, but there is a real risk that it will escape from containment and do tremendous damage.

By contrast, the risks of working with dangerous pathogens, and how to control those risks, are well understood. There are thorough, detailed and tested protocols for doing even the simplest kinds of biological research. Biosafety is not a political issue. A lab in Boston follows almost the same protocols as a lab in Beijing or Brussels.

None of that is true with respect to AI. Modern biological research protocols developed over nearly 140 years . Anthropic was founded in 2021. That’s why researchers and AI CEOs are sounding the alarm.

It is absolutely vital that we start figuring out how to regulate AI so that it doesn’t do immense, uncontrolled damage. But at the moment, we don’t even know where to begin. Asmodei’s proposal for slowing the pace of AI aren’t solutions. They are a way to buy us a little time to work out the solutions.

The good news is that laying the groundwork for regulating AI is the kind of thing that Congress was born to do. The House, the Senate or, preferably, both could create a well-staffed select committee tasked with producing a report on regulating AI. The public hearings might be the most consequential congressional hearings since Watergate, and the resulting report would become a springboard for crafting actual legislation and negotiating international agreements.

The bad news is that congressional Republicans won’t back such a plan because they are terrified that Trump will label them as “Conspiracy Theorists, Treasonists, Traitors, and Leakers.” It astounds me that Republican lawmakers prefer what may be a 10 percent chance of a world-wide catastrophe to the risk of the president saying something mean about them on social media, but here we are.

There’s actually something we ordinary citizens can do about all this. Keeping AI under control is far more important than Trump or any of our current political squabbles. We simply cannot wait another two years to start the process of creating a proper regulatory framework for the potential monster we are creating.

Republicans have had two years to tackle this problem, and they have refused . I realize this is a tough thing to ask if you are a Republican, but even Steve Bannon agrees that the uncontrolled development of AI is an imminent threat to humanity. The only path forward is to put Democrats in charge of Congress and get on with the job .

So this November, if you want to see AI and the tech oligarchs finally put on a leash, vote for a Democrat. If you believe there is a risk that AI may destroy humanity , there’s no other choice.

Chris Truax is an appellate attorney who served as Southern California chair for John McCain’s primary campaign in 2008.