When name recognition seems to perform thoughtful policy at the polls, political candidates constantly seek innovative ways to reach voters and increase turnout. Artificial intelligence is buzzy, but fears of deepfakes and misinformation have prompted everything from public pushback to outright bans. One congressional candidate, however, is embracing AI—and is deploying an AI version of himself to engage with voters.
“They can have a simple basic conversation with an AI me, [which can] answer some basic questions they have," Ohio Democratic congressional candidate Matthew Diemer said. “If there are more deep conversations they want to have, then there is a way to contact me and reach out to us via email, so I can call them back personally.”
Diemer is using an AI tool from London-based Civox, which provide chatbots that can effectively speak for a candidate as the candidate, interacting one-on-one (or AI-on-one) with many more constituents that would be possible for a mere, solo human.
“System like Civox allows me to put my voice out there to people,” Diemer said—and that would be over 730,000 citizens throughout the state.
Diemer, who was a periodic host on Decrypt's once-daily gm podcast, previously differentiated his candidacy through his support of crypto. Making AI only the latest emerging technology added to his toolbox.
The idea of AI-generated politicians may trigger memories of viral deepfakes of U.S. President Joe Biden, former President Donald Trump, and Pope Francis. And AI giants Meta and Anthropic have banned the use of their tools in election campaigns. But Diemer said Civox only uses his voice and its messages are based on a questionnaire provided to Diemer.
“It's no different than sending out blogs, emails, text messages, TikToks, or tweets,” Diemer said. “This is another way for people to interact with me and have more of a connection.”
Unlike sales calls that use lists of names and phone numbers bought by data brokers, Diemer emphasized that the only voters that the Civox AI will contact are those who have already signed up to receive updates from his campaign.
“We're not violating the law; we are calling the people that have already opted in for my list that want to talk to me that I want to reach back out to,” Diemer said. “I'm trying to make it more efficient so I can talk to them again. And if they really wanting to talk to me, they can give me a callback.”
Civox was co-founded in 2023 by Ilya Mouzykantskii and Adam Reis, who also founded Conversation Labs, which underpins the technology. The startup hopes to carve out a cottage industry of politics-focused AI models. In December, the campaign of Pennsylvania Congressional candidate Shamaine Daniels leveraged Civox to reach voters in the state’s 10th district.
“In one year, most Americans will have spoken with an AI being, whether they know it or not,” Reis said in a December press release. “We can either stand by and let bad actors use this once-in-a-species opportunity to unleash a flood of political misinformation, or we can get ahead of it with proactive regulation and education to realize its tremendous positive potential in government.”
Regulators, however, are attempting to stamp out AI-generated deepfakes, stepping up actions to curb the use of AI to influence voters and elections.
Earlier this month, a robocall that included an AI-generated deepfake of U.S. President Joe Biden went viral after attempting to dissuade New Hampshire voters from participating in the state’s primary. In response, the U.S. Federal Communications Commission officially made using AI-generated deep fakes in robocalls illegal.
While protecting voters from AI deepfakes is important, Diemer said regulators need to be careful not to create a monopoly on AI technology.
“The more that these technologies do exist and [regulators] limit the fashion that they're being allowed to be used by the average person or different businesses is siloing this technology,” Diemer said. “There will be companies that have it, but if it's not available to the people, that means that one company will be able to use it, abuse it, and influence it.”
Civox CEO Ilya Mouzykantskii agreed that people have a right to be concerned.
“The concern is very real, and as we get closer to November, you're going to start seeing companies that are competitive with Civox trying to deceive and misinform voters on both sides of the aisle,” Mouzykantskii told Decrypt. “At Civox, we take the high road.
“We are a fully managed service," he added. “We do not allow our customers to make deceitful calls, full stop.”
Mouzykantskii said Civox intends to exceed current AI regulatory requirements by disclosing at the start of each call that the caller is an artificial intelligence agent, a step he said that is not mandated by law in any state but that the company believes is ethical.
“The fact of the matter is the cat out of the bag—people are going to be making these calls,” Diemer said. “That's why we want to have these conversations because we want to be ahead of this... we want people to know that this is happening and [AI] is going to be used. We want to know that it can be used responsibly.”
Looking to establish best practices in the emerging AI industry, Mouzykantskii said Civox rejects requests and turns down potential clients who want to misrepresent its AI as human.
“AI is an incredibly powerful tool,” Mouzykantskii told Decrypt. “And we're just at the very tip of beginning to see all the ways in which AI will be applicable, particularly in politics,” he said.
Edited by Ryan Ozawa.