MICHEL MARTIN, HOST:
The State Department says it is investigating an incident last month where an AI-generated voice posing as Secretary of State Marco Rubio sent messages to at least three foreign ministers, a U.S. senator and a governor. NPR White House correspondent Deepa Shivaram reports.
DEEPA SHIVARAM, BYLINE: The deepfake of Rubio included at least two voicemails and a text message over the messaging app Signal. That's according to a July 3 memo sent out by the state department to staff members and embassies. It's not clear who the messages were sent to, though. The deepfake comes as more government officials are being targeted by generative AI scams. In May, President Trump's chief of staff, Susie Wiles, was a target of an AI deepfake. And last year, an AI-generated robocall in New Hampshire used former President Biden's voice to discourage Democrats from voting. Experts say more instances are likely.
IFEOMA AJUNWA: This capability of generative AI has already been weaponized for political purposes. And we just do not have the appropriate regulations in place to actually stop bad actors from using this technology in this way.
SHIVARAM: That's Ifeoma Ajunwa. She's a law professor at Emory University, and she says that any regulation that does come will have to be widespread.
AJUNWA: We just want to think about the way that this is a problem that's not containable and therefore actually requires some modicum of coordination between states or on a federal level - or even, frankly, between countries.
SHIVARAM: The Federal Communications Commission has ruled that using AI-generated voices in a robocall is illegal. But even that doesn't necessarily stop bad actors, says Greg Bohl, the head of data at Transaction Network Services. It's a company that helps protect consumers from scam calls. He says current regulations are just a speed bump for those creating deepfakes.
GREG BOHL: They can hide behind these larger systems and these networks, and it does take quite a bit of detective work to chase them down.
SHIVARAM: Bohl points out that more regulation would be helpful. But there are some steep challenges with coming up with a legal framework that punishes for crimes and also allows for innovation.
BOHL: How do you restrict bad actors by not restricting people to prevent the bad actors?
SHIVARAM: It's a challenge Congress has been working on for years. And despite bipartisan support for a bill, they still haven't come up with an answer.
Deepa Shivaram, NPR News.
(SOUNDBITE OF CHESSBOXER'S "I CAN'T TELL MY SECRET WEAPON") 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.
