With a Biden voice deepfake, AI wades into NH voter suppression

A robocall impersonating President Biden urged voters to skip the New Hampshire primary and “save your votes for November.” Experts say AI deepfakes such as this one can blur the concept of truth, posing a deeper challenge to democracy. 

President Biden speaks in the East Room of the White House on Jan. 19, 2024. The New Hampshire attorney general is investigating reports of robocalls that used AI to impersonate Mr. Biden's voice, spreading misinformation about the New Hampshire primary.

Evan Vucci/AP

January 23, 2024

The New Hampshire attorney general’s office on Jan. 22 said it was investigating reports of an apparent robocall that used artificial intelligence to mimic President Joe Biden’s voice and discourage voters in the state from coming to the polls during the Jan. 23 primary election.

Attorney General John Formella said the recorded message, which was sent to multiple voters on Sunday, appears to be an illegal attempt to disrupt and suppress voting. He said voters “should disregard the contents of this message entirely.”

A recording of the call reviewed by The Associated Press generates a voice similar to Mr. Biden’s and employs his often-used phrase, “What a bunch of malarkey.” It then tells the listener to “save your vote for the November election.”

In this Spanish town, capitalism actually works for the workers

“Voting this Tuesday only enables the Republicans in their quest to elect Donald Trump again,” the voice mimicking Mr. Biden says. “Your vote makes a difference in November, not this Tuesday.”

It is not true that voting in the Jan. 23 primary precludes voters from casting a ballot in November’s general election. Mr. Biden is not campaigning in New Hampshire and his name will not appear on Tuesday’s primary ballot after he elevated South Carolina to the lead-off position for the Democratic primaries, but his allies are running a write-in campaign for him in the state.

Gail Huntley, a Democrat in Hancock, New Hampshire, who plans to write in Mr. Biden’s name on Jan. 23, said she received the call at about 6:25 p.m. on Jan. 21.

She instantly recognized the voice as belonging to Mr. Biden but quickly realized it was a scam because what he was saying didn’t make sense. Initially, she figured his words were taken out of context.

“I didn’t think about it at the time that it wasn’t his real voice. That’s how convincing it was,” she said, adding that she is appalled but not surprised that AI-generated fakes like this are spreading in her state.

Historic Israeli desire to ‘go it alone’ is tested by Gaza and Iran

White House press secretary Karine Jean-Pierre confirmed Jan. 22 that the call “was indeed fake and not recorded by the president.” Mr. Biden’s campaign manager, Julie Chavez Rodriguez, said in a statement that the campaign is “actively discussing additional actions to take immediately.”

Mr. Biden and Vice President Kamala Harris will share the stage on Jan. 23 in Virginia as they campaign for abortion rights, a top issue for Democrats in an election expected to feature a rematch with Donald Trump, the former Republican president.

The apparent attempt at voter suppression using rapidly advancing generative AI technology is one example of what experts warn will make 2024 a year of unprecedented election disinformation around the world.

Generative AI deepfakes already have appeared in campaign ads in the 2024 presidential race, and the technology has been misused to spread misinformation in multiple elections across the globe over the past year, from Slovakia to Indonesia, and Taiwan.

“We have been concerned that generative AI would be weaponized in the upcoming election and we are seeing what is surely a sign of things to come,” said Hany Farid, an expert in digital forensics at the University of California, Berkeley, who reviewed the call recording and confirmed it is a relatively low-quality AI fake.

As AI technology improves, the federal government is still scrambling to address it. Congress has yet to pass legislation seeking to regulate the industry’s role in politics despite some bipartisan support. The Federal Election Commission is weighing public comments on a petition for it to regulate AI deepfakes in campaign ads.

Though the use of generative AI to influence elections is relatively new, “robocalls and dirty tricks go back a long ways,” said David Becker, a former U.S. Department of Justice attorney and election law expert who now leads the Center for Election Innovation and Research.

He said it’s hard to determine whether the main intent of the New Hampshire calls was to suppress voting or simply to “continue the process of getting Americans to untether themselves from fact and truth regarding our democracy.”

“They don’t need to convince us that what they’re saying, the lies they’re telling, are true,” he said. “They just need to convince us that there is no truth, that you can’t believe anything you’re told.”

Katie Dolan, a spokeswoman for the campaign of Rep. Dean Phillips of Minnesota, who is challenging Mr. Biden in the Democratic primary, said Mr. Phillips’ team was not involved and only found out about the deepfake attempt when a reporter called seeking comment.

“Any effort to discourage voters is disgraceful and an unacceptable affront to democracy,” Ms. Dolan said in a statement. “The potential use of AI to manipulate voters is deeply disturbing.”

This story was reported for The Associated Press.