JUANA SUMMERS, HOST:
AI is getting faster, better and, some would say, scarier. In fact, recently, a lot of voices from the tech industry have warned that AI is advancing too quickly and that it could cause great harm to humanity. And those voices, they include the heads of AI companies themselves. But in order to slow things down, there needs to be global cooperation, starting with the countries with the most advanced AI companies. Those are, of course, China and the United States. The leaders of those countries will try and find a way forward this week when Chinese leader Xi Jinping meets with President Trump here in Washington, D.C. But the last time this subject came up we didn't get very far. Seth Center led the U.S. State Department delegation in AI talks with China in 2024. Welcome to the program.
SETH CENTER: Thanks for having me.
SUMMERS: Thanks for being here. OK. You recently wrote a New York Times op-ed, and in part of it, you say, if the past is prologue, China will have little interest in anything beyond gaining a competitive advantage. So just walk us back a bit. What happened last time?
CENTER: The two presidents agreed in a summit that the United States and China should have a dialogue about AI safety and the possibility of cooperation to manage some of the risks. And we subsequently met in Geneva several months later to try and frame an agenda. And fundamentally what happened is the United States side took the mandate seriously and the Chinese side decided to play for geopolitical advantage. I think they were professional and courteous in their interactions with us, but they chose not to take the AI risk seriously and preferred to, I think, play what would be called their standard diplomatic playbook.
SUMMERS: I wonder if since those 2024 talks, have you seen anything that's changed to make you think that China might have a different posture this go around?
CENTER: First and foremost, there's a body of researchers within China who take these risks now, I think, more seriously. I think the dynamic has changed politically because China now has multiple capable models which give them a sense of confidence. And I think we have evidence that, publicly, China's leadership is speaking about safety and risk a little more openly than they did previously.
SUMMERS: As we mentioned, U.S. AI leaders have been just really vocal recently about the mounting risks. Do you think that makes a difference here?
CENTER: I think that China sees this situation a little bit differently. Imagine you're in their shoes. They know that there's growing skepticism in the United States of an AI buildout at the infrastructure level. And they see the leading proponents of the very best U.S. companies saying that our systems may be unsafe, so let's slow down. Now, if they take AI safety seriously, they will see this as a blinking warning for their own companies as well, which are only a few months behind ours. But from a strategic competition perspective, from Beijing's perspective, they may see this as a window of an opportunity to let the United States wrestle with the dilemmas of AI while they keep pushing their commercial models all around the world without talking directly about the safety of those systems.
SUMMERS: From my vantage point, there seems to be a bit of mixed messaging here. We've heard growing calls from U.S. lawmakers, industry leaders, asking for some sort of guardrails on AI. And yet, President Trump has rejected calls for constraints. He said that AI is the next industrial revolution, but larger and more impactful. What kind of impact could that sort of disconnect have?
CENTER: Well, I think both governments are struggling right now to find the right balance between strategic competition, technological competition, commercial competition, and acknowledging that there may be some risks that have to be confronted by governments. And the question in the U.S.-China relationship - and many of these debates about the role of government are playing out against the backdrop of a larger China competition - is if there's not going to be a mega treaty between the United States and China to slow down AI - and I don't think that's likely, and certainly nothing in the last couple of days makes it seem like this is likely to occur - could a small section of the AI conversation be carved out in which mutual interests between the U.S. and China could get them to cooperate in small areas to reduce risk very specifically where narrow national security risks are at stake.
SUMMERS: As we mentioned, you led the U.S. State Department delegation in AI talks with China two years ago. What advice would you give people who are in these conversations right now?
CENTER: One, manage expectations. I think it's very difficult to see a situation where China sees the overriding logic of its geopolitical ambitions being subordinated to AI risk concerns. Second, make sure there's a formula in place to test, press and verify China on any agreements that are made. And third, to - not to lose sight of the ball game, which is that the United States must lead in all facets of AI.
SUMMERS: Seth Center, former acting special envoy for critical and emerging technology is at the U.S. Department of State. He's now at Johns Hopkins University Seth, thank you.
CENTER: My pleasure.
(SOUNDBITE OF MINUTEMEN'S "COHESION") Transcript provided by NPR, Copyright NPR.
NPR transcripts are created on a rush deadline by an NPR contractor. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.
