AI Makes Political Texts More Effective and More Annoying
AI-powered platforms are training bots to sound like political candidates in text messages, holding personalized conversations with thousands of potential voters simultaneously as the 2026 midterm elections approach. While campaign strategists hail the technology as a revolutionary tool for voter engagement, critics warn it risks making an already intrusive channel even more annoying — and potentially deceptive.
The Rise of AI-Powered Political Texting
The political texting market greatly expanded in 2020 when COVID-19 forced campaigns to abandon door-to-door canvassing. Live phone banking declined as landlines disappeared, and social media algorithms increasingly controlled campaign reach. Text messaging offered a direct, algorithm-free channel to voters’ phones.
Now, generative AI is taking that channel to a new level. Companies like Akillion, Convos, and Vector Political are offering platforms that enable campaigns to deploy AI bots capable of holding nuanced, back-and-forth conversations with voters at scale.
Aaron Sheeks, CEO of Akillion, an AI platform that lets users run custom large language models, 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,” Sheeks told NPR. “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.”
How the Technology Works
In almost all cases, the first text message sent to voters is written and sent by a human, according to Eric Wilson, a Republican strategist and director of the Center for Campaign Innovation. The AI steps in when the recipient engages, taking over the conversation with personalized responses.
Tom Carroll, CEO of Convos, an AI-powered text messaging platform, described the approach: “What we’re offering is the greatest volunteer you’ve ever had. They’ll respond within 30 seconds, in any language, cutting directly to the question that the person is asking.”
Convos launched last year and helped with 10 political campaigns. This year, it is aiming to work with over 100 campaigns and has already reached roughly half that target, Carroll said.
Marty Santalucia, a partner at Vector Political, which focuses on generative AI texting, said his firm has sent 2.5 million text messages this year and generated 20,000 to 30,000 conversations. About 5-10% of recipients respond to texts, and 10-20% of those engage in 10 or more exchanges. “We’re listening at a scale that campaigns have never listened at before,” Santalucia told NPR.
The Partisan Divide
Experts say Republicans have been adapting to AI faster than Democrats in campaign messaging. Wilson, who trains Republican campaign strategists and candidates on AI adoption, said generative AI “helps campaigns do more with less.”
Wilson suggested the gap may reflect the fact that the two dominant political debates around AI — its environmental footprint and its impact on labor and unions — both cut against Democratic politics rather than Republican ones. “We just don’t have that on the right,” he said. “We’re focused on winning with the tools that we have.”
A Pew Research Center survey conducted in February 2026 found that Democrats are less confident than Republicans in the government’s ability to regulate AI effectively — 74% of Democrats expressed little to no confidence, compared with 61% of Republicans.
Ethical Concerns and the Disclosure Paradox
Not everyone in the political texting industry is on board with the AI revolution. Josh Justice, CEO of Peerly, a peer-to-peer texting platform, said he has ethical concerns with using generative AI to communicate with voters.
“I don’t think it’s ethical to use generative A.I. to communicate with voters,” Justice told NPR. “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.”
This tension — the disclosure paradox — lies at the heart of the debate. Requiring campaigns to disclose that voters are talking to an AI makes the practice more ethical, but it may also undermine the very effectiveness that makes AI texting attractive. Voters may be less willing to engage with a bot than a human volunteer.
Nathan Rifkin, co-CEO at Scale to Win, a tech company focused on grassroots organizing for progressives, warned of the risks. “Or you can lead AI chat bots to say some pretty horrific things,” Rifkin told NPR. “If that’s in the voice of the candidate, that can lead to some bad ends.”
A Patchwork of State Regulations
As of 2026, 30 states require AI disclosure in political ads, and at least 20 states have enacted laws targeting AI-generated deepfakes in political contexts, with another 15 considering similar legislation, according to The AI Lobby.
Campaigns in North Dakota and California must disclose if recipients are talking to virtual assistants in their first message. New Jersey may soon follow suit. However, there is no federal law specifically regulating AI in political campaigns as of July 2026.
The bipartisan Protect Elections from Deceptive AI Act (S. 3312), sponsored by Senators Klobuchar, Hawley, Coons, and Collins, would criminalize the distribution of materially deceptive AI-generated content about federal candidates within 60 days of an election. But the bill faces First Amendment concerns and has not yet passed.
Voter Annoyance and the Risk of Backlash
For voters like Stefanie Party, a 44-year-old Cleveland resident who moved back from Chile last year, the flood of political text messages — sometimes up to five a day — is already overwhelming. “You really can’t tell who they’re coming from,” Party told NPR. “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 nonprofit helping Democrats adopt digital marketing techniques, said data shows political text messaging “used to work till it got abused by overuse.” She argued that AI is best used to find new ways to reach people, not to rescue channels that voters already hate.
What’s Next
The 2026 midterm elections will serve as a live stress test for AI-powered political messaging. Key questions remain: Will the Protect Elections from Deceptive AI Act pass before November? How will voters react when they discover they’ve been conversing with AI rather than human campaign staff? And will the partisan gap in AI adoption affect election outcomes?
With 49% of U.S. adults now using AI chatbots — up from 33% in 2024, according to Pew — the technology’s role in campaigns is only likely to grow. But so too may the backlash. As one voter put it, the texts make her feel “super, super annoyed.” Whether AI makes that annoyance more tolerable or simply more pervasive may determine the future of political outreach.