Spoof phone calls and messages to lawmakers in which artificial intelligence programs imitate high-ranking officials are increasingly common on Capitol Hill.
The latest report that an AI impostor of Secretary of State Marco Rubio contacted several government officials has put senators on guard, and some are advocating for more AI constraints in light of the incident.
“I talk to him frequently, multiple times a day, sometimes,” Sen. James Risch said of Rubio. “And I always say, ‘Is this really you, Marco?’”
Risch, who chairs the Senate Foreign Relations Committee, said members on the panel have been getting an increasing number of phishing calls over the last six months. Recently, Risch himself got a call from someone claiming Speaker of the House Mike Johnson wanted to speak with him. Risch was suspicious, so he called Senate Majority Leader John Thune to make sure there was not an emergency. Risch said Thune told him the phishing calls had been happening to other people.
“I think the thing with Rubio has got a little new twist on it, using this AI sound bite,” Risch said. “They probably wouldn’t try it on me because I know Marco really well, and I could say a couple of things to know like that if this was Marco or not.”
The Washington Post reported Tuesday that an impostor used an AI-generated version of Rubio’s voice and writing style to message foreign ministers, a U.S. member of Congress and a governor, with the likely goal of manipulating these officials in order to gain information or account access.
Sen. Mike Rounds, a member of the Senate Select Committee on Intelligence, called the race to keep up with AI technology a “war.”
“We have to have better AI than the bad guys do,” Rounds told NOTUS. “This is a war, and this is a case where we cannot afford to lose in this battle. It’s an ongoing battle without an end in sight.”
“This is not new. It’s simply very, very good, and the technology is getting better all the time,” Rounds added. “We’ve always known about it. We’ve just always tried to keep our technology ahead of it in terms of detection, but this really does draw attention to it.”
Last year, an AI deepfake caller posing as a high-level Ukrainian official tried to pry sensitive information on the 2024 election from former Democratic Sen. Ben Cardin, who was then the head of the Senate Foreign Relations Committee. Since then, the capabilities of AI models have only grown.
Several lawmakers told NOTUS that the attempt to impersonate Rubio is an example of what’s on the horizon for AI. Sen. Mark Kelly, another Intelligence Committee member, called the situation “a national security concern.”
“We’re just going to see more of this as AI gets more effective,” Kelly said. “People have to be skeptical about where they’re getting their information from. Phone calls could create huge problems.”
Though President Donald Trump recently signed into law a bill that prohibits nonconsensual online posting of sexual imagery made with AI, senators used this situation as an opportunity to push for more legislative solutions.
Sen. Ron Wyden, another Intelligence Committee member, called the situation “outrageous” and mentioned his AI-regulation bill, the Algorithmic Accountability Act.
“It’s just a kind of measure of what challenges are going to look like down the road,” Wyden said.
The NO FAKES Act is another piece of legislation — reintroduced in the House and Senate earlier this year — intended to penalize the illicit use of an individual’s likeness created with AI deepfake technology.
However, additional AI regulation may be unlikely this Congress as the GOP has pushed for greater integration with it. Trump brought AI enthusiast David Sacks into the administration as a tech czar. And some congressional Republicans recently attempted, but ultimately failed, to prevent states from enacting AI regulation for 10 years.
Sen. Josh Hawley, who called for stripping that provision from the reconciliation bill, now wants Congress to work on measures to prevent more abuse using AI.
“We need to pass a regulatory framework that will give people rights to their name, image and likeness, that will protect copyrights, and that will require clear labeling of AI, and then allow private rights of action,” Hawley told NOTUS.