It begins with a woman named Noa—or maybe she calls herself Ruth, or Ella. Her profile picture shows a smiling young woman in a Tel Aviv café, a scarf around her neck, a laptop open on the wooden table. She has just moved to Israel from France. She posts in Facebook groups for olim asking where to buy a good challah, complaining about the price of cottage cheese, sharing a heartbreaking story about being rejected by her family because she became Jewish. In one group, a kind person sends her a long private message with advice on health insurance, and she thanks them warmly. She is lovely. She is also not real. Her face was generated by a machine, her biography assembled from fragments of other people’s lives, her “kindness” engineered by a foreign influence operation. According to a disturbing investigation carried out by Haaretz, this is the new face of election interference in Israel. It is not only about fake accounts or bots, but about a whole architecture of fabricated people—fake olim, AI hostages—designed to weaken the country from within. The goal of these foreign networks is not to help any particular party win. It is to make the democratic contest itself seem absurd, to flood the public square with enough poison that no one can tell what is real and what is not. If the election is chaotic, if voters are disgusted, if the result is questioned and hated by the losing side, their operation has succeeded.
Why invent new immigrants? Because Israel is a country of immigrants. “Oleh” is a sacred category. The state proudly encourages aliyah, and any new arrival is treated with special affection. There is a deep cultural impulse to welcome the stranger, to help them find a job, a community, a shred of normalcy. Foreign influence networks know this. The fake olim exploit the one thing Israel cannot easily defend: its open door to the Jewish world. They do not arrive at Ben Gurion airport; they appear in WhatsApp groups and Telegram channels. Their Hebrew is broken, they ask innocent questions about army service, they express bewilderment at Israeli politics. That vulnerability makes them trusted. It gives them license to say things a native Israeli would be challenged for. A fake oleh can casually mention that “they say the prime minister steals everything” or that “the army is killing children in Gaza” and then, when confronted, hide behind ignorance: “I’m new, I don’t understand.” The report traces dozens of such accounts with uncanny similarities: they register at the same time, use the same profile-picture generator, share the same syntactical errors. They are spread across different political tribes. Some pose as liberal secular immigrants horrified by religious extremism; others pose as religious immigrants who are afraid to wear kippot on the street. They are like a virus that can attach to any cell in the body politic. Their effect is to amplify and polarize every existing dispute inside Israeli society—between left and right, Ashkenazi and Mizrahi, religious and secular, Jew and Arab. They make a country that is already contested even more divided, because a society arguing with itself is less able to respond to external threats. And because they are not voters, they have no stake in the country. They are internal enemies that do not exist, ghosts who take up residence in the public’s imagination.
Then there are the AI hostages. The phrase itself is unbearable. After October 7, when Hamas abducted around 250 people, the images of hostages became a central part of Israeli civic grief. Pictures of the kidnapped children, women, and elderly are everywhere—on billboards, on supermarket doors, inside prayers. Foreign networks have learned how to weaponize this grief. They have created hostage stories that never happened, videos of “abducted” people who are not on any official list, who have no families searching for them, no name in the state’s records. An AI-generated face can be manipulated to cry, to plead, to tremble, to speak in Hebrew with the accent of a kibbutznik. A template video can be made of a young woman with a gun pointed at her head, saying her name is “Maya” and asking the government to surrender to Hamas. But Maya does not exist. In some cases, the report explains, real hostage photographs are blended with artificial elements, taken from old social media profiles, altered so that family members cannot immediately prove they are fake. The result is not just disinformation. It is emotional warfare. Every person watching is forced to ask: is this real? If we cannot weep for a real victim without wondering if their face is synthetic, then our empathy itself becomes disabled. This is the deeper evil of the “AI hostage.” It turns the daily horror of the hostage crisis into a psychological contest. It also makes the viewer a hostage of his own doubt. We are not merely fooled; we are forced to hold uncertainty in our hearts while massive human tragedy unfolds. That is exactly what foreign networks want: a society whose compassion is so exhausted, so worn down by confusion, that it cannot mobilize around its own dead and taken.
Who is behind this? The Haaretz investigation points to coordinated networks linked, in the professional judgment of intelligence researchers and digital forensic analysts, to foreign state actors, especially Iran and Russia. These countries have no favorite candidate in the Israeli election; their objective is chaos for its own sake. Tehran wants to destabilize and weaken a regional enemy. Moscow wants to distract the West and show that all democracies are equally rotten. The playbook is familiar from America, Ukraine, France, Germany: create fake accounts, hire armies of trolls, amplify extreme voices, invent scandal, push divisive hashtags, and then deny everything. But in Israel, the operation is tuned specifically to local frequencies. It uses Hebrew slang and Jewish liturgy. It sends fake messages that quote the Bible and the Talmud, that invoke the memory of the Shoah, that mention the names of cities like Sderot and Ashkelon. It knows that Israelis are right and left; it inflames both. One day it promotes an angry protest against Netanyahu; the next it supports annexation and says Arabs are animals. The goal is to make every Israeli hate another Israeli a little more. It may also try to suppress voting: posting fake messages about “election day is moved” or “voting by mail is not necessary” in Telegram groups aimed at particular audiences. It works to create the impression that the election is rigged, that ballots are untrustworthy, that the entire political system is theater. If the winner can be challenged by the losing side as illegitimate, the networks have succeeded. They are not trying to elect someone; they are trying to make Israel unelectable. Democracy survives when losing a vote still means accepting the process; foreign manipulation wants to destroy that acceptance, to make the result itself seem like an injustice.
The report is grim, but it tries to describe not only the threat but the defense. Social media platforms, for years, have struggled to police disinformation in Hebrew because it is a small language. Many moderators do not recognize cultural cues. The number of reporters, researchers, and analysts working on this issue is tiny compared with the armies of bots. Yet Israeli civil society is fighting back. Independent fact-checkers painstakingly trace reverse-image searches and metadata. University researchers train algorithms to detect synthetic faces. Veterans of the cyber-intelligence world, now ordinary citizens, volunteer to spread awareness. The most important defense, the report suggests, is ordinary awareness. We do not need to be paranoid; we need to be attentive. When a stranger enters a group and immediately stirs controversy, screen them. When a video causes heart-stopping grief, check the source. Before sharing, ask: who made this, what is their history, are other sources confirming it? The phrase “trust but verify” is not just a Cold War proverb but a survival skill. Social media companies, too, must be held accountable—but the article warns that waiting for them may be too late. The gap between the speed of foreign manipulation and the speed of institutional response is huge. The networks are fast; vaccines are slow. That is why the human layer matters. Neighbors, friends, community leaders, rabbis, teachers, parents: they can build a civic immune system by teaching people to pause before reacting. The goal is not to silence doubt but to prevent manipulation from hijacking our emotions. A foreigner cannot steal our election if we refuse to let a stranger tell us how to feel.
In the end, the most human response is also the most radical. Democracy is not only a voting machine. It is a shared imagination: we believe in the reality of each other’s suffering, the legitimacy of each other’s voices, and the possibility of making decisions together. Foreign networks understand this. They attack the imagination itself by saturating it with artificial faces, fake olim, AI hostages. They want us to look at a neighbor and wonder: are you real? We must resist that suspicion. Not by naively trusting everyone, but by building a deeper trust through actual relationships, in person, in our streets, in our buildings, at Shabbat tables and planning committees. A bot does not sit beside you at a neighbor’s funeral; it does not bring you soup when you are sick; it does not stand in the rain to vote. The answer to fake olim is real olim—and real Israelis, of all backgrounds—who know each other by name. The answer to AI hostages is to hold our real hostages close: to remember their names, to pray for their release, to fight for them with a love that cannot be reproduced by a machine. The Haaretz investigation arrives in time for the election, but its deeper message goes beyond politics. We are in a war of reality. Be a human, not a network. Protect the truth by making truth something we share, not something we secretly own.