The article argues that the next Israeli election may be the country’s first true AI election, and that its main danger will not be a single dramatic deepfake but a wider campaign of confusion. It points to Slovakia’s 2023 election as an early warning: two days before voting, a fake recording spread on social media in which pro-European candidate Michal Šimečka seemed to admit, in a conversation with a journalist, that he had rigged the election. The voice was AI-generated, and because the clip spread during the media blackout period, it could not be quickly debunked. By the time the fraud was exposed, Šimečka had already lost.
The article says political actors are now racing to use cheaper, quieter AI tools to collect data, segment audiences and tailor messages. Instead of persuasion, the goal is to overwhelm voters with contradictory content until they stop trying to determine what is true. It cites RAND’s 2016 term “flood of lies” and says AI has made such propaganda nearly free to produce. Low-quality clips are not necessarily mistakes, they help normalize the idea that everything is fake. A quote from Yoni Orich, a Netanyahu adviser, is used to illustrate the logic of volume: after criticism of materials he circulated, he wrote, “4 videos of 20 seconds and Eisenkot wants a debate... wait, there are 400 more.”
The piece also highlights the “liar’s dividend,” the growing ability of politicians to dismiss damaging evidence by blaming AI. It says a 2024 study found that claims of fake content can boost support among a politician’s base more than silence or apology. The second prong of the threat is personalized persuasion: according to the Israel Internet Association, about 25% of Israelis are considering asking a language model whom to vote for. Campaigns are also expected to use GEO, or Generative Engine Optimization, seeding the web with targeted content so chatbots repeat the desired message. Experiments show that speaking with a language model can shift voter attitudes more than traditional ads.
The article warns that this is especially dangerous in Israel, where voters are already exhausted after a prolonged war and a post-traumatic period. It says the old anchors of shared fact, journalism, courts and visible evidence, are weakening, and AI has even undermined the last one, video proof. The author concludes that the key defense is recognizing the shape of an influence operation, since covert persuasion works only while it remains hidden.
