Israel's October 27 election is the first national contest where AI-generated content has moved from novelty to standard campaign infrastructure. According to Ze Be AI, a monitoring project at the Open University of Israel, more than 14% of the 44,000 social media posts tracked since late July across Facebook, Instagram, X, and TikTok used AI in some form. Netanyahu's Likud party alone had produced more than 1,800 AI-generated posts by mid-October — roughly a quarter of all Likud social media material analyzed. The content ranges from the absurd to the incendiary. Netanyahu's account posted a deepfake of rival Gadi Eisenkot — whose son was killed in the Israel-Hamas war — running past his son to embrace an Arab politician, a video designed to weaponize anti-Arab sentiment and viewed more than 1 million times. Finance Minister Bezalel Smotrich posted a photorealistic AI video depicting the attorney general and Supreme Court justices being detained at a protest. National Security Minister Itamar Ben-Gvir shared an AI-generated video of Palestinians on a conveyor belt entering what resembled a concentration camp, drawing condemnation from Germany's Foreign Ministry. Beyond content creation, AI is being deployed for voter contact at scale. Researchers documented a four-minute robocall from a bot identifying itself as 'Danielle' from a polling institute — conversational enough to field follow-up questions, persistent enough to steer every exchange back to voting preferences, and opaque enough to refuse five consecutive requests to identify its commissioner. The capacity constraint that once limited phone campaigns — real humans with real hours — has been removed. The audience is being manufactured alongside the content. Scooper, an Israeli media monitoring firm, found more than 1,000 new accounts appearing daily on X and immediately following Netanyahu. About 70% of accounts sharing Netanyahu's posts on X are based outside Israel and exhibit bot signatures: round-the-clock posting, recent creation dates, missing profile pictures. FakeReporter, an Israeli watchdog, identified a separate bot network of over 100 accounts linked to Ben-Gvir's campaign, some using AI-generated images and identities. Perhaps most consequential is the emerging front of generative engine optimization — creating web content designed to be ingested by AI chatbots and shape their answers. The Israeli government has already contracted a US firm, documented under Foreign Agents Registration Act filings, to build websites feeding LLM systems with pro-government content. About a quarter of Israeli voters said they would consider using AI tools to inform their voting decisions, according to the Israel Internet Association. The same infrastructure built for international information operations is now being turned inward. Regulatory infrastructure is nearly absent. Israel's Central Elections Committee issued a labeling mandate for AI content in July, but enforcement has been sporadic — Smotrich was ordered to remove one unlabeled video, while the broader flood continues. There is no AI Act, no Digital Services Act, and privacy law is outdated by the admission of researchers tracking the problem. The asymmetry is structural: AI lowers the cost of producing propaganda and manufacturing audiences faster than institutions can detect or counter it. Tehilla Shwartz Altshuler of the Israel Democracy Institute frames Israel as a test case for global democracies, given Israelis' early adoption of new technology. The implications extend well beyond one election cycle. If the machinery works — if deepfakes shift votes, bots manufacture consensus, and GEO poisons the information well that voters increasingly rely on — every subsequent election worldwide will face the same playbook, deployed by whoever can afford the compute.
