A robocall seemingly using artificial intelligence to mimic Joe Biden’s voice has been used to discourage people from voting in a primary election in the US.
The New Hampshire attorney general’s office said it was investigating after the recorded message was sent to several voters on Sunday ahead of Tuesday’s primary election.
The call begins with the US president’s frequently used phrase: “What a bunch of malarkey.”
A voice similar to Mr Biden’s says: “It’s important that you save your vote for the November election.
“Voting this Tuesday only enables the Republicans in their quest to elect Donald Trump again. Your vote makes a difference in November, not this Tuesday.”
It is untrue that voting in the primary on Tuesday precludes voters from casting a ballot in November’s general election.
While Mr Biden is not campaigning in New Hampshire and his name will not appear on the primary ballot – due to the president choosing South Carolina as the party’s first official primary – his allies are running a campaign urging voters to write in his name.
The White House confirmed Mr Biden did not record the call and said it highlights the challenges such emerging technologies present, especially ahead of the November presidential election.
“The president has been clear that there are risks associated with deepfakes. Fake images and misinformation can be exacerbated by emerging technologies,” White House press secretary Karine Jean-Pierre said.
Concerns over use of AI in election interference
New Hampshire attorney general John Formella said the recorded message appears to be an illegal attempt to disrupt and suppress voting, adding that voters “should disregard the contents of this message entirely”.
The proliferation of generative AI, which can create text, photos and videos in response to prompts, has been met with excitement about its potential as well as fears it could make jobs obsolete and be used to interfere in elections.
The technology has been used to spread misinformation in elections from Slovakia to Taiwan – with deepfake audio purporting to capture Labour leader Sir Keir Starmer abusing party staffers posted on the first day of the Labour Party conference.
Read more:
The world is about to have its biggest election year ever – and AI experts are sounding the alarm
Revealed: How many jobs will be affected by AI
Please use Chrome browser for a more accessible video player
PM hails ‘landmark’ AI agreement
The ‘biggest threat to democracy’?
“This is not the first time that AI technology has been used to mimic a politician and it will not be the last,” warned Arshin Adib-Moghaddam, professor in Global Thought and Comparative Philosophies at SOAS, University of London.
AI systems are becoming “increasingly adept at creating incredibly realistic deepfakes”, he told Sky News, with it being possible to create a simple deepfake, in which a voiceover is matched to a video, “in a minute by anyone online and for free”.
Prof Adib-Moghaddam said the central argument of his book, Is Artificial Intelligence Racist?, is “that AI technologies used for nefarious purposes are the biggest threat to democracy and human security in general”.
“Since the scandal of Cambridge Analytica, we know that AI systems can easily translate our personal data into incredibly targeted and minute propaganda, that makes Orwell’s 1984 seem like a harmless parody,” he added. “In this way, the criminal use of AI technology is emerging as the biggest threat to our cherished freedoms.”