Former Secretary of State Anthony Blinken Talks Rapid AI Advancement

SPEAKER_1 (0:02)

Bloomberg Audio Studios, podcasts, radio, news.

Joe Mathieu (0:07)

Mr. Secretary, it's great to have you back on Bloomberg TV and radio. We meet at a precarious time today. The war with Iran is stretching on. Oil prices are at their highest. We've seen in some time, certainly since the open of this invasion, interest rates are at their highest in years, and it's all taking place against the backdrop of an argument over an emerging technology that some believe could bring the end of humanity.

It's just another Tuesday. But I'd like to start there with you in this AI debate that's underway right now in Washington. The people building the most advanced AI models are saying it's time to slow down. The president of the United States says, we have to win. Whoever wins AI wins. Is the president right or the CEOs?

Antony Blinken (0:50)

These shouldn't be irreconcilable positions because here's the reality, Joe.

The statement that we heard the other day from Dario Modai, from Sam Altman, from Elon Musk, saying we've got to do something, that should be a big flashing red light. This is like the wizard's telling you something's rotten in Oz. We ought to listen and pay attention when we hear that. And the reality is this. This technology is moving at a speed that we can't even begin to wrap our arms around, never mind get ahead of. We're seeing a doubling in capacity on the frontier platforms. Every several months, Moore's Law is out the window when it comes to computers.

So the necessity of doing something to put some governing rails on this, I think is increasingly evident to pretty much everyone, although the president seems to disagree. Look, we regulate the airplanes flying in the skies. We regulate the nuclear power plants producing power for us. We regulate the stuff that we put into our bodies through pharmaceuticals. AI has the potential both for incredible good, but also for terrible consequences if it's not properly guarded. I'm convinced we can do this in a way that doesn't set back our innovation, that doesn't set us back in the race with China. By the way, it's very good that the Treasury Secretary is meeting with his Chinese counterpart.

Joe Mathieu (2:10)

Yeah.

Antony Blinken (2:11)

I suspect this will be front and center on the agenda.

Here's the thing. Do you think that China wants an AI that is, from the perspective of an autocratic state, an AI running a muck either in the hands of people that it doesn't control or in its own hands in ways it can't control? There is a way of bringing us together to see if we can't find some agreement.

Joe Mathieu (2:30)

Well, should there be an AI safety framework on the table when the President meets with Xi next week?

Antony Blinken (2:35)

Look, first of all, we need that at home.

Joe Mathieu (2:37)

So that doesn't give the win to China?

Antony Blinken (2:39)

It doesn't give... No, I don't think it gives the win to China. I think there's a shared interest for all of our differences, for all of the very profound mutual suspicions. The Chinese are going to be very suspicious of us, just as we're rightly very suspicious of them. So it doesn't make this easy, but we know from decades of arms control during the Cold War of the Soviet Union, these kinds of things are possible as hard as they are. Again, I don't want to minimize the difficulty, but I want to emphasize the necessity. And by the way, it's not just...

It's the governing principles, it's trust, it's safety, it's security. It's in some way getting a partnership between the private sector and government to embed experts, perhaps, to look at every model before it goes out, but to do it in a way that doesn't hold back innovation. We need to get all the companies on board, we need to get China on board, we need to get the world on board, and finally, our communities on board. We need a new social compact. If you're building a data center, we've got to make sure that the electricity rates aren't going to kill the community, that they're not going to lose out on water, that jobs are not going to go in the wrong direction, that there's upskilling, and the community benefits. If we do all of those things, I think we have a chance to harness this for good and to try to minimize the negative.

Joe Mathieu (3:48)

You mentioned Scott Besson's meeting with his counterpart from China this weekend. The channel for AI is going through Treasury. Why is this not being led by the State Department?