Gallego warns foreign adversaries are using American AI against the US, calls for guardrails
"We just, for example, learned that the Houthis were using Claude AI to figure out how to calibrate their missiles to better attack U.S. ships," Democratic U.S. Sen. Ruben Gallego said.
Watch: Exclusive: Sen. Ruben Gallego on AI threats, regulation and Senate oversight | 12 mins
America’s adversaries are using U.S.-developed artificial intelligence against us, according to U.S. Sen. Ruben Gallego, who wants stronger safeguards and mandatory reporting of foreign theft or misuse.
The Arizona Democrat cited a recent report during a Thursday interview with KTAR News 92.3 FM’s Arizona’s Morning News.
“We just, for example, learned that the Houthis were using Claude AI to figure out how to calibrate their missiles to better attack U.S. ships,” Gallego said.
How are foreign adversaries using American AI technology?
According to Reuters, Anthropic said threat actors in northern Yemen used Claude to develop, test and troubleshoot software for missile and rocket guidance systems.
“We know Iranians are also doing the same thing,” Gallego said. “Claude and a couple of AI companies caught them, and were able to cut them off, but there are a lot of things that are leaking out.”
While he welcomed Anthropic’s disclosure, Gallego said reporting such incidents shouldn’t be voluntary.
“Our U.S. companies also need to tighten up,” he said.
What are Ruben Gallego’s AI goals?
Gallego called for stronger cybersecurity and ongoing monitoring to prevent espionage and theft across both open- and closed-source AI. His particular concern is protecting “frontier models,” the most advanced AI systems available.
“These are the ones that are very dangerous when put in the wrong hands. If a country or an entity steals one of these models, there is no requirement for these AI companies to tell the U.S. government, ‘Hey, somebody just broke in and stole one of our very, very, very, very, very important models,'” Gallego said.
For Gallego, notifying the government is part of the industry’s responsibility to protect national security.
“They share the responsibility — the core responsibility and responsibility as Patriots — to tell us when foreign actors are using our own tech against us,” he said.






Comments
0 Comments