Charly Triballeau/AFP via Getty Images
Stay up to date with our Politics newsletter, sent weekly.
AI-powered platforms are training bots to sound like political candidates in text messages, holding personalized conversations with thousands of potential voters simultaneously. The bots are also gathering data, learning what each voter wants from their representatives and using that information to shape future campaign messaging.
Aaron Sheeks, the CEO of Akillion, an AI platform that lets people run their own Large Language Models or bots, said many of his current clients are running for political office.
“Our goal is to put the microphone back in the hand of the voter,” said Sheeks. “We’re giving agencies and political campaigns the ability to have a trained AI employee that can go back and forth and answer questions on police reform or education or tax changes.”
Some in the broader political text messaging industry say generative AI’s ability to answer voter’s questions and gather data about their concerns is going to be revolutionary for campaigns; others say political text messages are a limited — and annoying — tool and using AI won’t improve it. While it’s difficult to ascertain how many campaigns are using generative AI to text voters this election cycle, experts say Republicans have been adapting to AI faster than Democrats.
“My belief is that this is going to make campaigns more interactive, more responsive and more personalized,” said Eric Wilson, a Republican strategist and the director of the Center for Campaign Innovation, a nonprofit that encourages conservative campaigns to adopt new technology. He said that generative AI “helps campaigns do more with less.”
In almost all cases the first text message sent to voters is written and sent by a human, Wilson said. The AI steps in when the recipient engages.
The era of the long political text message is over, said Tom Carroll, the CEO of Convos, an AI-powered text messaging platform. Convos guides campaigns and their bots to say a sentence, introduce themselves and then ask a question to initiate a conversation.
“What we’re offering is the greatest volunteer you’ve ever had,” said Carroll. “They’ll respond within 30 seconds, in any language, cutting directly to the question that the person is asking.”
Convos launched last year, Carroll said, and helped with 10 political campaigns. This year they’re aiming to work with over 100 campaigns; so far, they’ve hit roughly half of that target.
Marty Santalucia, a partner at Vector Political, which is focused on generative-AI texting, said bots excel in engaging voters and “in some cases, we have people talk to our agent for hours.” About 5-10% of people respond to texts and about 10-20% of those engage in 10 or more texts.
“We’ve sent two and a half million text messages this year and had over 20,000 to 30,000 conversations,” said Santalucia. “We’re listening at a scale that campaigns have never listened at before.”
The rise of the campaign text message
The market of political texting greatly expanded in 2020 as candidates were struggling to connect with voters on a grassroots level, said Josh Justice, the CEO of Peerly, a peer-to-peer texting platform. Live phone callers and phone banks diminished as landlines died out. Door-to-door canvassing grew outdated as people grew suspicious of answering the door for strangers. And on social media, the relationship between candidates and voters is owned by the tech platform.
Justice thinks every single House or Senate campaign this year will send out text messages because it’s one of the few ways to reach voters directly on a mass scale. It stays on voters’ phones, and it doesn’t have to compete with the algorithm for attention.
Justice and others who work for traditional political texting companies said they have ethical concerns with generative AI political text messaging. AI can be used for data analysis or for guiding volunteers, he said, but campaigns should immediately reveal to voters that they are talking to persuasive bots.
“I don’t think it’s ethical to use generative A.I. to communicate with voters,” said Justice. “You can put a disclaimer on there, and that’s going to make it a lot better. But that defeats the purpose of what everybody started out doing.”
He is particularly concerned about campaigns using this while laws regulating AI are still being put into place. Campaigns in North Dakota and California have to tell recipients if they are talking to virtual assistants in their first message. Campaigns in New Jersey may soon have to disclose when they use generative AI to provide election-related information to voters.
Nathan Rifkin, co-CEO at Scale to Win, a tech company that does grassroots organizing and fundraising for progressives, said the risk of using generative AI outweighs the benefits, including the chatbot giving false information.
“Or you can lead AI chat bots to say some pretty horrific things,” said Rifkin. “If that’s in the voice of the candidate, that can lead to some bad ends.”
Tech companies selling generative AI text messages to political candidates say their clients aren’t keen on going public. Some of that is because candidates don’t want to share their “secret sauce” said Marty Santalucia from Vector Political. He also admits another element is that it’s “very muddy in terms of where public perception is going to fall on this tool.”
A Pew Research Center survey showed Democrats are less confident than Republicans in the government’s ability to regulate AI effectively. Democratic campaigns are more hesitant to try out new technology, while Republican campaigns are more energetic about experimentation, Santalucia said.
Wilson, who trains Republican campaign strategists and candidates on how to adopt AI, thinks the difference between the campaigns could be because the two political debates around AI —– its environmental footprint and its impact on labor and unions — skew against Democratic politics.
“We just don’t have that on the right,” said Wilson. “We’re focused on winning with the tools that we have.”
The platform could be the problem
Stefanie Party, 44, moved back from Chile to Cleveland, Ohio, last year and that’s when the political messages, often clickbaity in nature and sometimes up to 5 a day, began. They make her feel “super, super annoyed,” she said.
“You really can’t tell who they’re coming from,” Party said. “Even if I’m talking to AI that claims to be giving me good information or personalized information, I really have no idea who’s on the other side of that.”
Jessica Alter, co-founder and chair of Tech for Campaigns, a political nonprofit helping Democrats adopt data and digital marketing techniques, said data shows that political text messaging used to work till it got abused by overuse.
While texting can still be useful in increasing voter turnout and generative AI texting might help, Alter said AI is best used to find new and measurable ways to connect with people.
“I think AI is not best used to, like, rescue 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 find new ways to reach people.”
