Political campaigns and their vendors are increasingly experimenting with generative-AI tools that can reply to voters by text in near real time, answer questions on issues such as education and taxes, and collect data about what recipients care about—an approach some operatives say Republicans have adopted more quickly than Democrats.
AI-powered texting platforms are training bots to sound like political candidates and to hold personalized conversations with large numbers of potential voters at once. The systems are typically triggered only after an initial text written and sent by a human, according to Eric Wilson, a Republican strategist who leads the nonprofit Center for Campaign Innovation.
“My belief is that this is going to make campaigns more interactive, more responsive and more personalized,” Wilson said, adding that generative AI “helps campaigns do more with less.”
Some vendors say the technology can also shorten and streamline the standard campaign text. Tom Carroll, the CEO of Convos, said campaigns using his company’s product are encouraged to send a brief introduction and a question to start an exchange, and he said the bots can reply quickly in multiple languages.
Carroll said Convos launched in 2025 and worked with about 10 campaigns that year, and that the company is aiming to work with more than 100 campaigns in 2026.
Another firm, Vector Political, said its generative-AI texting has been used at scale this year. Marty Santalucia, a partner at Vector Political, said the company had sent about 2.5 million text messages in 2026 and logged roughly 20,000 to 30,000 conversations, with some exchanges lasting for hours.
The spread of AI in political texting has prompted warnings from parts of the traditional political texting and organizing industry. Josh Justice, the CEO of Peerly, said campaigns should disclose to recipients when they are interacting with an AI system, and progressive organizing technology executive Nathan Rifkin of Scale to Win said chatbots can deliver false information or be prompted into harmful responses—risks that could be amplified if the messages are presented in a candidate’s voice.
Regulation is evolving unevenly across states. North Dakota enacted a 2025 law requiring certain political communications that use AI to include an artificial-intelligence disclosure, though the statute is focused on content that visually or audibly impersonates a human and does not broadly cover routine text generation. In New Jersey, legislation introduced in 2026 would require disclosures when generative-AI chatbots provide voters with election-related information.
Even without AI, campaign texting has become a major feature of modern elections. Industry executives told NPR that the political texting market expanded sharply during the 2020 cycle as campaigns sought alternatives to traditional field tactics. But heavy volumes of messages have also led to widespread voter irritation, and some Democrats and Republicans alike question whether adding AI will make the channel more effective—or simply more prolific.
