In a recent video widely circulated online, Israel’s top justices and its attorney general are seen being dragged by security guards from a chaotic anti-government protest that never took place.

The 33 seconds of emotionally-charged footage was originally posted by Finance Minister Bezalel Smotrich, a fierce opponent of the judiciary, and triggered an immediate backlash. Critics said it incited violence, while supporters said it was justified because they accuse the leaders of Israel’s legal system of a left-wing plot to stymie the government’s policies.

But the depicted scenes are fake — no such protest took place and the video carries a tell-tale watermark to identify content made using artificial intelligence. Smotrich’s post was just the latest example of a deluge of AI-crafted distortions that are saturating the Israeli information environment just weeks before the country’s most consequential elections in decades.

“This is Israel’s first generative AI election,” said Anat Ben-David, who leads a project called “This is AI” at Israel’s Open University, which has tracked the surge in AI-generated campaign material over recent weeks. Other experts worry that the country’s electoral regulatory system is not fit to handle the threat.

Since starting to monitor the trend in late July, Ben-David’s team has found ubiquitous use of AI across the political spectrum, although about 45% has been produced by Prime Minister Benjamin Netanyahu’s campaign. Netanyahu heads a religious-rightist coalition that includes Smotrich, a Jewish settler on the West Bank and proponent of a one-state solution that would annex Palestinian territories, as well as ultra-Orthodox parties and other groups.

Representatives for Netanyahu and Smotrich’s offices didn’t immediately respond to requests for comment.

Candidates who oppose Netanyahu’s re-election in the Oct. 27 vote, including the left-wing Democrats party headed by Yair Golan, have taken to using squishy toys as a recurring AI prop to demonstrate the prime minister’s pliability at the hands of his partners. Smaller, cash-strapped parties have also been drawn to AI, Ben-David’s team found, given its ability to create buzz without a big budget.

The candidates use AI to edit takes, add jingles and overlay subtitles, but have attracted controversy when they deploy the technology to add a manipulative political context. It’s a tool being increasingly used around the world, having already disrupted elections from Europe to India, and has emerged as an influential factor in campaigning ahead of next month’s vote in Brazil.

Israel’s elections will be the country’s first since the deadly Hamas attacks on southern Israel on Oct. 7, 2023, which triggered the war in Gaza, making them particularly emotionally charged. AI is being used to revive old controversies and sharpen internet echo chambers.

In previous rounds, Netanyahu, the country’s longest serving leader, has resorted to demonizing Arab Israelis and associated parties to re-enforce his reputation as “Mr. Security.” That constituent, making up about 20% of Israel’s population, will likely play a pivotal role in this year’s vote.

“Elections are not a civil war, and neither are AI videos,” President Isaac Herzog, who holds a largely ceremonial role, said last week following the spread of Smotrich’s video clip. He called for the preservation of national unity in the face of the growing phenomenon.

A Channel 13 poll on Thursday indicated that Netanyahu’s bloc has recovered some ground and is projected to win around 54 seats, against 52 for the opposition, which has congregated under a potential coalition led by Gadi Eisenkot’s Yashar party. Both are short of the 61-seat parliamentary 120-seat majority, meaning negotiations will likely follow about adding parties to form a government.

In an election of slim margins, Israeli candidates are leveraging AI to manipulate voters’ emotions to the point of persuasion.

On July 20, Netanyahu set the tone with an AI generated ad that depicted his top contender, Eisenkot, about to embrace a young man who resembles his late son – killed while serving in Gaza – but instead hugs the Arab Islamist leader Mansour Abbas.

The sleek video was set at sunset, in a field with wildflowers. It was taken down soon after controversy broke for defying the sacrosanct status usually granted to parents whose children have been killed during Israeli military service.

Days later, the Israeli Central Elections Committee rushed through a law requiring AI watermarks be embedded across at least 5% of the image, to identify that the depicted people, events or public monuments are fakes.

The goal, said Dean Livne, the committee’s acting director, is to balance “the integrity of the elections” against “free speech.”

Only a handful of complaints have come to his office since the new law was enacted, he said. He also dismissed suggestions that the watermark is ineffective and said his office is preparing an around the clock “war room” for election day to ensure that illegal posts be “taken down within minutes.”

Amit Zac, an assistant professor at Hebrew University in Jerusalem, said the regulatory bodies are outdated and not fit to combat the scale, speed and sophistication of the AI campaign that is playing out.

“They say, ‘take that post down,’ but on the internet that’s meaningless,” he said.

Netanyahu’s AI ad against Eisenkot was removed weeks ago, as was an AI ad posted by National Security Minister Itamar Ben Gvir, which boasted of having denied food to Palestinian prisoners. But deleted videos remain searchable and many observers expect that they will continue to proliferate.

“The more the legal implementation of these policies fail,” said Zac, “the more the space becomes a Wild West.”

—With assistance from Chris Miller.

©2026 Bloomberg L.P. Visit bloomberg.com. Distributed by Tribune Content Agency, LLC.