As election campaigns ramp up across the US, the state’s attorney general’s office in New Hampshire is investigating an AI-related robocall incident.
This AI-generated call impersonated President Joe Biden, intending to dissuade voters from participating in the upcoming primary election.
Attorney General John Formella described the incident as a seemingly illegal act aimed at disrupting and suppressing voter turnout. He emphasized, “Voters should disregard the contents of this message entirely.”
The call, distributed to several voters on Sunday, used AI to closely replicate President Biden’s voice, even incorporating his commonly used phrase, “What a bunch of malarkey.”
The message misguidedly advised listeners to “save your vote for the November election,” falsely claiming that voting in the primary would increase Donald Trump’s chance of winning.
This is incorrect, as participation in the primary does not affect the ability to vote in the general election.
Interestingly, President Biden is not campaigning in New Hampshire and is not on the primary ballot. However, a write-in campaign in his favor is ongoing in the state.
The source of these calls remains unknown, but they were falsely presented as originating from Kathy Sullivan’s personal cellphone number.
Sullivan, a former state Democratic Party chair involved with the pro-Biden super-PAC Granite for America, called the incident “outright election interference” and a clear attempt to harass her and other voters. She reported, “This call links back to my personal cell phone number without my permission.”
Julie Chavez Rodriguez, Biden’s campaign manager, forcefully responded, stating that the campaign is considering immediate additional actions.
She condemned the spread of disinformation aimed at suppressing voting, asserting, “Spreading disinformation to suppress voting and deliberately undermine free and fair elections will not stand.”
This incident highlights a growing concern about using generative AI technology for election disinformation. Hany Farid, a digital forensics expert at the University of California, Berkeley, confirmed the recording as a low-quality AI fake. He expressed concerns about AI being weaponized in elections, particularly if more effective tech is deployed.
In response, Katie Dolan, a spokeswoman for Rep. Dean Phillips, who is running against Biden in the Democratic primary, disavowed any involvement and stressed that such efforts to discourage voters are disgraceful and deeply disturbing.
The Trump campaign, also implicated due to the nature of the message, stated they had no involvement but declined to comment further.
This follows Pennsylvania congressional candidate Shamaine Daniels’s use of AI-augmented robocalls. Daniels has employed an AI robocaller named Ashley, developed by Civox in collaboration with Conversation Labs, to efficiently reach a large number of potential supporters.
Ashley enables two-way conversations so voters can ask questions about Daniels’ campaign and policies.
Ilya Mouzykantskii, the co-founder of Civox, indicated Ashley is “significantly less expensive than human-paid phone bankers.”
The use of AI in election campaigning is clearly risky, and there have been several examples of when deep fakes might’ve manipulated people’s behavior.
