AI disinformation and foreign interference in Israeli polls

In the lead-up to Israel’s October 27 election, artificial intelligence has become a central tool for political manipulation and personal attacks. Researchers at Golden Owl identified a coordinated disinformation network, assessed to be linked to Iran, involving nearly 120,000 social media accounts. This network utilized AI-generated content and deepfakes to amplify divisive narratives regarding Prime Minister Benjamin Netanyahu, the October 7 failures, and military draft exemptions.

As the campaign progressed, domestic political actors increasingly employed AI-generated imagery and videos to target opponents. With polling showing a near-tie between the Netanyahu bloc and the center-right bloc led by Gadi Eisenkot, both sides have used AI to attack reputations. This includes Eisenkot using AI depictions of Iranian and Hamas figures to criticize Netanyahu, and Netanyahu using AI-generated content that appeared to manipulate scenes involving a young man resembling Eisenkot’s deceased son.

Israeli security agencies, including the police and the Shin Bet, have since moved to disrupt networks on X/Twitter suspected of attempting to influence the Knesset elections. In coordination with the State Attorney’s Office and the Central Elections Committee, authorities removed numerous accounts to prevent undue influence. Investigations by the Shin Bet and the Institute for National Security Studies suggest these efforts may involve foreign actors from Iran, Russia, and Islamist organizations operating from Turkey. Some accounts reportedly posed as Israeli citizens to spread misinformation regarding the voting process and claims that election results would be stolen.