ICYMI: “Senator Coons urges Congress to act on AI risks” in Bloomberg interview

WILMINGTON, Del. - In case you missed it, U.S. Senator Chris Coons (D-Del.) joined Bloomberg This Weekend yesterday to discuss the growing risks of artificial intelligence and the need for Congress to establish guardrails to protect against the biggest risks that the technology can pose. The interview follows Anthropic CEO Dario Amodei's call for AI companies to slow the development of their most advanced models, along with warnings about the potential for catastrophic harm, and the resignation of Anthropic researcher Jacob Coxon, who said Anthropic and OpenAI were "gambling with our lives."

During the interview, Senator Coons called for Congress to work together to establish enforceable guardrails. He noted that the U.S. must work with allies and adversaries alike to mitigate the growing risks. He also highlighted the AI Whistleblower Protection Act, his bipartisan legislation with Senator Chuck Grassley to protect AI whistleblowers.

Bloomberg This Weekend: Senator Chris Coons on AI and the need for guardrails

Senator Coons: The essay by Dario Amodei says out loud what many of us have known for months, if not years. Especially after the [Hugging] Face hack that happened recently. These models are doing things that their designers didn't expect, can't predict, and aren't able to fully control. And the level of risk for us as a species, as a nation, is gradually accelerating. I have a bill in the Judiciary Committee with Senator Grassley that we are going to try and get passed soon that would protect AI whistleblowers. Because frankly, the main reason we know about these steadily increasing risks is because of whistleblowers. They're taking career risks, even personal risks, by stepping out and saying this. Now that we've got agreement from Elon Musk and Sam Altman and Dario Amodei, we need agreement from Republicans and Democrats in Congress to put aside our partisan pre-election bickering and put in place some real guardrails before AI goes off the rails.

…

I think all of us are concerned about our competition with China, or we should be, and the ways in which AI is accelerating developments in everything, not just positive developments in, for example, finding new treatments and cures for diseases, but developments in the possibility of AI engineering a deadly bio-weapon, AI being used for hypersonic missiles, for cracking codes, for taking down GPS. There are lots of malign applications for AI in warfare. The drones that are being fired at Ukrainian civilians every night by Russia are becoming faster and stronger because Russia and Iran are using AI to improve their targeting. So, yes, there is a real concern on my part and many of my colleagues.

…

I also, frankly, think that this is something that has to be on the agenda when President Trump meets with Xi Jinping. This is as much a threat to China as it is to the United States, and so, even though our competition is fierce, even though our political systems are fundamentally different, I'll remind you: in the height of the Cold War, when the Soviet Union and the United States were locked in a decades-long conflict, where we threatened each other's existence with nuclear weapons, we were able to negotiate arms control agreements to reduce the risk we would destroy humanity. This moment requires us to do the same: to look up from our domestic concerns or our partisan concerns and work to find a way that we can come up with some minimal, reasonable guardrails to slow the rapidly accelerating pace of AI development and deployment.