In this file photo, a man using a smartphone passes through an arched window at Grand Central Terminal on March 24, 2025 in New York City.
Charlie Tribalew/AFP via Getty Images
hide caption
toggle caption
Charlie Tribalew/AFP via Getty Images
Stay informed with our weekly politics newsletter.
The AI-powered platform trains bots to sound like political candidates in text messages, engaging in personalized conversations with thousands of potential voters simultaneously. Bots also collect data to learn what each voter wants from their representatives and use that information to shape future campaign messages.
Aaron Sheeks, CEO of Akillion, an AI platform that allows you to run your own large language models and bots, said many of his current customers are running for political office.
“Our goal is to put the microphone back in the hands of voters,” Sheeks said. “We’re empowering law enforcement agencies and political campaigns to have trained AI employees who can go back and forth and answer questions about police reform, education, and tax reform.”
Some in the broader political text messaging industry say generative AI’s ability to answer voters’ questions and collect data about their concerns will be revolutionary for campaigns. Some say political texting is a limited and annoying tool that will not be improved by the use of AI. It’s difficult to know how many campaigns are using generative AI to send text messages to voters this election cycle, but experts say Republicans are adapting to AI faster than Democrats.
“I believe this will make campaigns more interactive, more responsive, more personalized,” said Eric Wilson, a Republican strategist and director of the Campaign Innovation Center, a nonprofit that encourages conservative campaigns to embrace new technology. He said generative AI “can help campaigns achieve more with less effort.”
In most cases, the first text message sent to voters is written and sent by a human, Wilson said. The AI intervenes when the recipient engages.
Tom Carroll, CEO of Convos, an AI-powered text messaging platform, said the days of long political text messages are over. Convos guides campaigns and their bots to say sentences, introduce themselves, and then ask questions to start a conversation.
“We have the best volunteers ever,” Carroll said. “They answer the other person’s question directly in any language within 30 seconds.”
Convos was launched last year and has supported 10 political campaigns, Carroll said. This year, they aim to work on over 100 campaigns. So far, we have achieved about half of our goal.
Marty Santalucia, a partner at Vector Political, which focuses on generated AI text messages, said bots are great at engaging voters and “in some cases, people end up talking to agents for hours.” About 5-10% of people respond to text messages, and about 10-20% of people engage in 10 or more text messages.
“This year, we’ve sent 2.5 million text messages and had over 20,000 to 30,000 conversations,” Santalucia said. “We are listening on a scale that campaigns have never listened before.”
Increased campaign text messages
The market for political messages expanded significantly in 2020 as candidates struggled to connect with voters at the grassroots level, said Josh Justice, CEO of Peary, a peer-to-peer texting platform. As landlines were phased out, live callers and phone banks declined. Door-to-door solicitation became obsolete as people became suspicious of answering the door for strangers. And on social media, technology platforms control the relationship between candidates and voters.
Justice believes every House or Senate campaign will send text messages this year because it’s one of the few ways to directly reach voters on a large scale. It’s stored on voters’ phones and doesn’t have to compete with algorithms for their attention.
Justice and others who work for traditional political texting companies said there are ethical concerns about generative AI political texting. He said AI can be used to analyze data and coach volunteers, but campaigns need to immediately make it clear to voters that they are talking to a persuasive bot.
“I do not think it is ethical to use generative AI to communicate with voters,” the judge said. “If you put a disclaimer in there, it would be a lot better. But it defeats the purpose of what everyone was trying to do in the first place.”
He is particularly concerned about campaigns using AI while laws regulating it are still in place. Campaigns in North Dakota and California require recipients to tell whether they are talking to a virtual assistant in the first message. New Jersey campaigns may soon need to clarify when to use generative AI to provide election-related information to voters.
Nathan Rifkin, co-CEO of Scale to Win, a tech company that does grassroots organizing and fundraising for progressives, said the risks of using generative AI, such as chatbots giving out false information, outweigh the benefits.
“Alternatively, you can make an AI chatbot say some pretty scary things,” Rifkin said. “If it’s in the candidate’s voice, that can lead to some bad outcomes.”
A tech company that sells generated AI text messages to political candidates says its customers are reluctant to go public. Vector Political’s Marty Santalucia said part of the reason is that candidates don’t want to share their “secret sauce.” He also acknowledged that another factor is that “the public perception of this tool is very murky in terms of where it’s going.”
A Pew Research Center survey found that Democrats are less confident than Republicans in the government’s ability to effectively regulate AI. Santalucia said Democrats are reluctant to try new technology, while Republicans are more willing to experiment.
Wilson, who teaches Republican campaign strategists and candidates how to implement AI, believes the difference between the two campaigns may be due to the fact that two political debates around AI – its environmental impact and its impact on labor unions – skew toward Democratic politics.
“We just don’t have a right winger,” Wilson said. “We are focused on winning with the tools we have.”
There may be a problem with the platform
When Stephanie Partey, 44, returned to Cleveland, Ohio from Chile last year, she began sending out political messages, often clickbait, sometimes up to five a day. They make her “very, very frustrated,” she says.
“I really don’t know who they’re coming from,” Partey said. “You may be having a conversation with an AI that claims to give you good information or personalized information, but you have no idea who is on the other side.”
Jessica Alter, co-founder and president of Tech for Campaigns, a political nonprofit that helps Democrats implement data and digital marketing technology, said data shows political texting was working until overuse led to abuse.
While texting can still help increase voter turnout, and generative AI texting can help, Alter said AI is best used to find new and measurable ways to connect with people.
“I don’t think AI is really used to rescuing channels that people already hate,” Alter said, referring to text messages from political campaigns. “It’s best used to find new ways to do things and new ways to reach people.”
