Ahead of the 2026 midterms, political campaigns are deploying AI chatbots trained to sound like the candidate in text message conversations with voters. The first message is human-written. When a voter replies, the AI takes over — responding within 30 seconds, in multiple languages, adapting to what each voter said. Platforms like Convos, Akillion, and Vector Political are selling this as a way to hold thousands of personalized voter conversations that campaigns couldn't staff with people. The FEC declined to require AI disclosure in political messaging in 2024, and only two states — North Dakota and California — currently require campaigns to say in the first text that they're talking to a virtual assistant.

1. Campaigns Can Finally Have Real Conversations With Voters (Eric Wilson, Center for Campaign Innovation; Tom Carroll, Convos; Aaron Sheeks, Akillion)

These platforms say AI texting gives every voter a personalized back-and-forth that campaigns couldn't afford to staff before.

AI texting bots are already running in dozens of campaigns this cycle. Convos helped 10 campaigns in 2025 and is targeting more than 100 campaigns this cycle, with roughly 50 secured so far. Vector Political has already sent 2.5 million texts and had 20,000–30,000 full conversations. Tom Carroll, Convos' CEO, calls the AI "the greatest volunteer you've ever had."

Voters actually talk back — and campaigns are listening. Aaron Sheeks, CEO of Akillion, says the goal is to "put the microphone back in the hand of the voter." The AI captures what each voter says they want and feeds that data back to the campaign in near real-time. About 10–20% of people who respond to political texts at all exchange 10 or more messages.

Republicans are adopting this faster than Democrats. Eric Wilson, executive director of the Center for Campaign Innovation, trains Republican campaign strategists on AI and says it will make campaigns "more interactive, more responsive and more personalized." Democrats are slower to adopt. On the left, AI carries political liabilities — its environmental footprint and its effects on union labor — that make it a harder sell.

2. But Voters Don't Know They're Talking to a Bot (Josh Justice, Peerly; Nathan Rifkin, Scale to Win)

These firms are direct competitors to the AI platforms — and they say campaigns are impersonating a candidate with AI without telling voters.

Voters can't tell they're talking to a bot. Stefanie Party, a voter in Ohio, said: "You really can't tell who they're coming from. Even if I'm talking to AI that claims to be giving me good information, I really have no idea who's on the other side of that." An NBC News poll found 89% of voters say people should be clearly told when content is AI-generated.

Adding a disclaimer would help — and would also kill the product. Josh Justice, CEO of Peerly (a traditional texting platform), says he doesn't think using generative AI to communicate with voters is ethical. He adds that a disclaimer would help, "but that defeats the purpose of what everybody started out doing."

An AI chatbot speaking as the candidate can go badly wrong. Nathan Rifkin, co-CEO of Scale to Win — a progressive campaign tech firm — says the risk outweighs the benefit: "You can lead AI chat bots to say some pretty horrific things. If that's in the voice of the candidate, that can lead to some bad ends." He's particularly worried campaigns are doing this while lawmakers are still writing the rules.

3. Still, AI Can't Fix a Channel Most Voters Already Resent (Jessica Alter, Tech for Campaigns)

Tech for Campaigns supports AI in campaigns — just not in text messaging, which most voters already find annoying.

Political text messaging is already a channel most voters ignore. Only 5–10% of voters respond to political texts at all. Jessica Alter, co-founder of Tech for Campaigns, says "AI is not best used to rescue channels that people already hate."

AI should open new channels, not polish a disliked one. Alter says campaigns should use AI to find new ways to reach voters — not to perfect a method they already tune out. Making a resented channel more personalized doesn't fix why voters resent it.

Where This Lands

The campaigns running AI bots say they're giving voters real personalized contact — something campaigns have never managed before at scale. Critics inside the same industry say campaigns are letting voters assume they're talking to the candidate when they're talking to a bot. North Dakota and California require disclosure in the first message; the FEC has declined to weigh in nationally. Alter argues that even if campaigns started disclosing the bot, they'd still be pouring resources into a tool most voters don't want.

Sources