By
Steven Kramer, a political consultant who commissioned a deepfake of President Joe Biden's voice that encouraged New Hampshire voters to stay home during the Democratic primary, was indicted on Thursday.
In separate announcements Thursday, New Hampshire’s attorney general charged Kramer with 26 counts, while the FCC fined him $6 million for “scam calls he set up to defraud voters” in violation of a federal Caller ID law. “New Hampshire remains committed to ensuring that our elections remain free from unlawful interference and our investigation into this matter remains ongoing,” New Hampshire Attorney General John Formella said in a statement. “I hope that our respective enforcement actions send a strong deterrent signal to anyone who might consider interfering with elections, whether through the use of artificial intelligence or otherwise.”
Kramer, who was employed by long-shot Democratic presidential candidate Rep. Dean Phillips’ campaign at the time, was charged with 13 felony counts of voter suppression and 13 misdemeanor counts of impersonation of a candidate. Kramer himself and Phillips’ now-defunct campaign have both said the campaign was unaware of the robocall.
Kramer has also been fined $6 million by the Federal Communications Commission.
In public comments, Kramer has gone with a whistleblower-adjacent defense, claiming he commissioned the deepfake as a way to prompt regulation of artificial intelligence tools. He told NBC, “This is a way for me to make a difference, and I have.”
That is arguably one way to do that, but people have managed to draw attention to the dangers of deepfakes in creative ways that — unlike Kramer’s approach — haven’t ended in criminal charges. Which isn't to say that it wasn't effective: In response to the scandal, the FCC swiftly moved to declare robocalls made using A.I. to be illegal.
One thing is clear: This technology is in need of stricter rules to protect against A.I.-powered voter suppression and manipulation. Deepfake audio and video tools are widely accessible, and they’re only going to become more capable. Studies in recent years have shown most people already have trouble identifying deepfake audio and video, meaning the potential for nefarious, A.I.-powered political activity is a clear and present danger.