Israel's 2026 Election Campaigns Deploy AI Bots and Robocalls

Israel's campaigns embrace unregulated AI arsenal

Israel's October 2026 election is unfolding in a dramatically different technological landscape than the country's last general election in November 2022, which took place just before ChatGPT launched. Political campaigns now deploy a vast array of AI-powered tools to identify, target, and persuade voters—with virtually no regulatory framework to govern their use.

According to a Haaretz investigation, Israeli parties are purchasing bot services that systematically vote in online polls to create false momentum, using AI-generated robocalls that personalize messages to individual voters, and employing advertising intelligence (ADINT) technologies originally developed to track Iranian nuclear scientists. Campaign operatives and technology vendors confirmed these practices are widespread in the current cycle.

Poll manipulation at industrial scale

One of the most brazen new services involves fleets of bots that scan social media and the wider internet for open polls, then systematically vote for their candidate. These bots can manipulate unregulated surveys that often draw hundreds of thousands of participants on platforms like Instagram.

While such polls lack statistical credibility, they serve strategic purposes: generating perceived momentum, damaging rivals' images, or falsely reassuring voters that a party near Israel's 3.25 percent electoral threshold faces no danger of failing to qualify for the Knesset. Israel's Central Elections Committee recently required media outlets to submit data on statistically valid polls within 24 hours, but this directive does not cover open internet polls—leaving a significant loophole.

From espionage tech to voter targeting

Campaigns are also leveraging ADINT capabilities that link supposedly anonymous digital advertising identifiers to phone numbers and other personal data. These technologies, once reserved for intelligence operations, now allow parties to track smartphone users, identify their political leanings, and target them with tailored messages.

Parties integrate this geolocation and behavioral data with voter registries from the Interior Ministry, creating detailed profiles enriched by AI analysis of massive commercial databases. Some campaigns have even built custom voter relationship management systems using AI models like Anthropic's Claude.

AI avatars and volunteer armies

Beyond automated bots, campaigns deploy sophisticated avatar networks—fake accounts maintained partially by humans that infiltrate closed groups and amplify party messages while appearing indistinguishable from real users. At least two such networks are active in the current election, serving candidates across the political spectrum.

Parties also recruit thousands of real supporters who receive daily briefings with AI-drafted messages, images, and videos to share, creating the appearance of organic grassroots activity. AI agents scan the internet for discussion hubs where these volunteers are then directed to engage.

The most evolved technology involves AI-powered robocalls that personalize greetings and responses to individual voters, drawing from preprepared answer sets to simulate genuine conversation. Campaigns are also optimizing content so AI chatbots like ChatGPT reflect positive information about their candidates when users ask political questions.

Why it matters

Israel's election laws, drafted in the 1950s and only recently updated to address deepfakes, have not kept pace with AI-powered campaign technologies. No government body regulates the tools described in this investigation, creating what one privacy attorney called "a mass casualty event for privacy." The gap is particularly concerning because these same technologies can be weaponized after polls close to generate stolen-election narratives and undermine public trust in results. The regulatory vacuum appears to benefit incumbents: as one former Likud campaign manager stated in 2019, "We have no red lines. The goal is to win—we'll do damage control later."

These details were first reported by Haaretz in an investigation by Omer Benjakob.