In Concord, New Hampshire, a political consultant has found himself at the center of a legal and political maelstrom. Steven Kramer, the consultant in question, is facing serious charges in connection with distributing AI-generated robocalls that impersonated the voice of former President Joe Biden. This act, purportedly designed to expose the risks posed by artificial intelligence, has evolved into a significant court case that raises questions about electoral integrity and the role of AI in modern campaigns.
In Belknap County Superior Court, where the trial commenced recently, Assistant Attorney General Brendan O’Donnell characterized the case as a “brazen attack” on the New Hampshire presidential primary for the 2024 election, which historically holds the distinction of being the nation’s first. Prosecutors argue that the message conveyed by the robocalls, delivered just two days ahead of the primary, was not just an AI mimicry of President Biden’s recognizable phrase, “What a bunch of malarkey,” but a misleading directive implying voters should refrain from voting until November’s general elections.
Kramer, who faces a potential decades-long prison sentence upon a conviction for charges of voter suppression and candidate impersonation, contends that his actions were misinterpreted. Instead of attempting to influence election outcomes, he asserts his objective was to alert the public to the perils AI poses. The consultant admitted to commissioning the $150 recording from a New Orleans magician, intending to spark a conversation about technology’s looming threats on democracy.
Prior to the trial, prosecutors sought to exclude discussions around whether the primary could be considered legitimate, given its changed status with the Democratic National Committee (DNC). While the DNC, at President Biden’s behest, initially attempted to remove New Hampshire from its prime spot in the electoral calendar, they later withdrew their threat to disqualify state delegates from the national convention. Though Biden did not actively participate or campaign, he was declared victorious through a write-in campaign.
Judge Elizabeth Leonard deemed relevant Kramer’s interpretation of these DNC actions. However, she maintained that legally, the state conducted its primary on January 23, 2024. Jurors will be briefed on this ruling but won’t be obligated to accept it as definitive in their considerations.
During the opening arguments, Kramer’s defense, led by attorney Thomas Reid, posited that the robocall was merely Kramer’s critique of the DNC’s maneuvers, describing them as an actual affront to New Hampshire’s primary. Reid argued that Kramer’s actions arose from skepticism about the primary’s authenticity, rather than malicious intent.
The defense countered claims of candidate impersonation by pointing out that the calls neither mentioned Biden by name nor directed voters to abstain from voting. Contradictorily, the prosecution presented several witnesses who perceived the message as discouraging participation. Theodore Bosen, a retired lawyer, expressed his dismay at the perceived manipulation attempt, labeling it as “horrific.”
Despite contentious court proceedings, witnesses unanimously testified that the messages did not deter them from participating in the elections. They maintained their understanding that primary votes did not exclude them from participating in the general election, although their awareness of the DNC’s actions varied.
The prosecution accused Kramer of attempting to obscure his involvement by allegedly using his father’s account for payments and inventing client identities when communicating with companies responsible for the call distribution. O’Donnell argued these actions evidenced Kramer’s awareness of wrongdoing and his intent of evasion.
This trial unfolds amidst a shifting national discourse on AI use in campaigns. Kramer has already incurred a $6 million fine from the Federal Communications Commission for his actions, with its payment status unclear. Although the initial political landscape had been inclined towards tight AI regulation, recent developments suggest potential deregulatory trends, with Congress deliberating on forbidding local AI regulations for the upcoming decade.