AI-generated voices in robocalls can deceive voters. The FCC just made them illegal

By ALI SWENSON

NEW YORK — The Federal Communications Commission on Thursday outlawed robocalls that contain voices generated by artificial intelligence, a decision that sends a clear message that exploiting the technology to scam people and mislead voters won’t be tolerated.

The unanimous ruling targets robocalls made with AI voice-cloning tools under the Telephone Consumer Protection Act, a 1991 law restricting junk calls that use artificial and prerecorded voice messages.

The announcement comes as New Hampshire authorities are advancing their investigation into AI-generated robocalls that mimicked President Joe Biden’s voice to discourage people from voting in the state’s first-in-the-nation primary last month.

Effective immediately, the regulation empowers the FCC to fine companies that use AI voices in their calls or block the service providers that carry them. It also opens the door for call recipients to file lawsuits and gives state attorneys general a new mechanism to crack down on violators, according to the FCC.

The agency’s chairwoman, Jessica Rosenworcel, said bad actors have been using AI-generated voices in robocalls to misinform voters, impersonate celebrities and extort family members.

RELATED: Fake Biden robocalls in New Hampshire traced to Texas company, criminal investigation underway

“It seems like something from the far-off future, but this threat is already here,” Rosenworcel told The Associated Press on Wednesday as the commission was considering the regulations. “All of us could be on the receiving end of these faked calls, so that’s why we felt the time to act was now.”

Under the consumer protection law, telemarketers generally cannot use automated dialers or artificial or prerecorded voice messages to call cellphones, and they cannot make such calls to landlines without prior written consent from the call recipient.

The new ruling classifies AI-generated voices in robocalls as “artificial” and thus enforceable by the same standards, the FCC said.

Those who break the law can face steep fines, with a maximum of more than $23,000 per call, the FCC said. The agency has previously used the consumer law to clamp down on robocallers interfering in elections, including imposing a $5 million fine on two conservative hoaxers for falsely warning people in predominantly Black areas that voting by mail could heighten their risk of arrest, debt collection and forced vaccination.

The law also gives call recipients the right to take legal action and potentially recover up to $1,500 in damages for each unwanted call.

Josh Lawson, director of AI and democracy at the Aspen Institute, said even with the FCC’s ruling, voters should prepare themselves for personalized spam to target them by phone, text and social media.

“The true dark hats tend to disregard the stakes and they know what they’re doing is unlawful,” he said. “We have to understand that bad actors are going to continue to rattle the cages and push the limits.”

Kathleen Carley, a Carnegie Mellon professor who specializes in computational disinformation, said that in order to detect AI abuse of voice technology, one needs to be able to clearly identify that the audio was AI generated.

That is possible now, she said, “because the technology for generating these calls has existed for awhile. It’s well understood and it makes standard mistakes. But that technology will get better.”

Sophisticated generative AI tools, from voice-cloning software to image generators, already are in use in elections in the U.S. and around the world.

Last year, as the U.S. presidential race got underway, several campaign advertisements used AI-generated audio or imagery, and some candidates experimented with using AI chatbots to communicate with voters.

Bipartisan efforts in Congress have sought to regulate AI in political campaigns, but no federal legislation has passed, with the general election nine months away.

Rep. Yvette Clarke, who introduced legislation to regulate AI in politics, lauded the FCC for its ruling but said now Congress needs to act.

“I believe Democrats and Republicans can agree that AI-generated content used to deceive people is a bad thing, and we need to work together to help folks have the tools necessary to help discern what’s real and what isn’t,” said Clarke, D-N.Y.

The AI-generated robocalls that sought to influence New Hampshire’s Jan. 23 primary election used a voice similar to Biden’s, employed his often-used phrase, “What a bunch of malarkey” and falsely suggested that voting in the primary would preclude voters from casting a ballot in November.

“New Hampshire had a taste of how AI can be used inappropriately in the election process,” New Hampshire Secretary of State David Scanlan said. “It is certainly appropriate to try and get our arms around the use and the enforcement so that we’re not misleading the voting population in a way that could harm our elections.”

The state’s attorney general, John Formella, said Tuesday that investigators had identified the Texas-based Life Corp. and its owner, Walter Monk as the source of the calls, which went to thousands of state residents, mostly registered Democrats. He said the calls were transmitted by another Texas-based company, Lingo Telecom.

According to the FCC, both Lingo Telecom and Life Corp. have been investigated for illegal robocalls in the past.

Lingo Telecom said in a statement Tuesday that it “acted immediately” to help with the investigation into the robocalls impersonating Biden. The company said it “had no involvement whatsoever in the production of the call content.”

A man who answered the business line for Life Corp. declined to comment Thursday.

Associated Press writers Christina A. Cassidy in Washington and Frank Bajak in Boston contributed to this report.

The Associated Press receives support from several private foundations to enhance its explanatory coverage of elections and democracy. See more about AP’s democracy initiative here. The AP is solely responsible for all content.