Israel’s parliamentary election scheduled for 27 October 2026 has become a testing ground for novel forms of political manipulation. Beyond traditional rallies and door-to-door canvassing, two parallel campaigns are drawing intense scrutiny: a series of counterfeit posters aimed at the United Arab List and a flood of AI-generated videos that blur the line between reality and fabrication. Both tactics appear designed to exploit existing social fissures and to sway undecided voters in the final weeks before polls open.
Fabricated signs target United Arab List and Mansour Abbas
In the days leading up to the vote, dozens of billboards and street signs surfaced across Israel displaying the image of United Arab List leader Mansour Abbas alongside a rainbow flag and the national flag. The graphic, which never received official approval from the party, was quickly denounced as a false campaign orchestrated by right-wing factions seeking to tarnish the Arab party’s reputation. The United Arab List asserted that the smear was intended to link the party to LGBTQ activism, a contentious issue among many Arab Israeli voters. Party officials emphasized confidence in “the awareness of the Arab public and its ability to deal with the smears,” suggesting that the community would recognize and reject the manufactured narrative.
AI-driven deepfakes turn the election into a digital battlefield
A separate wave of manipulation leveraged artificial intelligence to produce a striking 19-second clip posted on the official X account of Prime Minister Benjamin Netanyahu. The video showed Likud candidate Gadi Eisenkot sprinting through a meadow, briefly passing a figure resembling his late son, before embracing Mansour Abbas in a display meant to provoke anti-Arab sentiment. The clip amassed over one million views within hours, sparking outrage from veterans of the 2023 Israel-Hamas war who noted the painful reference to Eisenkot’s loss. Research by the Open University of Israel’s Ze Be AI project identified more than 1,800 AI-enhanced posts released by Likud by mid-October, accounting for roughly a quarter of the party’s social-media output. Across all parties, the study recorded 44,000 political posts, with over 14 % containing some AI component—ranging from simple image filters to full-blown deepfake videos.
While Likud leads in sheer volume, fringe parties exhibit the highest proportion of AI content relative to their total output. Left-wing rivals such as the Democrats and the Yashar party also deployed AI, though on a markedly smaller scale. Notable examples include a video of Democrat leader Yair Golan rotating under a ceiling fan, each spin altering his attire to reference military service and popular protest slogans, and a photorealistic clip by Finance Minister Bezalel Smotrich that portrayed the attorney general and Supreme Court justices being detained at an anti-government protest. National Security Minister Itamar Ben-Gvir circulated an AI-generated scene of Palestinians on a conveyor belt resembling a concentration-camp setting, prompting condemnation from multiple foreign ministries.
Regulatory response and the rise of AI-enabled robocalls
In July, Israel’s Central Elections Committee introduced a rule obligating campaigns to label any AI-generated or manipulated material. Beyond this labeling requirement, the legal framework remains thin: Israel lacks a comprehensive AI act, a digital services directive, and its privacy legislation is widely regarded as outdated. Consequently, the election arena has become a showcase for emerging technologies that would be difficult to deploy under stricter European regulations. One novel development is the use of AI-powered robocalls that sound conversational rather than mechanical. Disinformation researcher Nitsan Yasur captured a four-minute call in which a synthetic voice identified itself as “Danielle, a pollster,” proceeding to ask the respondent about voting preferences while repeatedly refusing to disclose its sponsor. The call illustrates how AI-driven outreach can bypass human labor constraints and potentially shape voter behavior at scale.
Election officials have taken limited action so far. In July, the committee’s chairman ordered Smotrich to remove an unlabeled AI video deemed misleading. However, enforcement remains inconsistent, leaving room for further exploitation as the campaign intensifies. The combined effect of fabricated signage and sophisticated AI content points to a rapidly evolving battlefield where traditional political messaging coexists with digital deception, raising pressing questions about accountability, media literacy, and the future shape of democratic contests.
