Meta, OpenAI, and investigative reporters uncovered a covert influence operation linked to Israel’s Ministry of Diaspora Affairs that used fake Western identities, AI-generated content, and anti-Muslim narratives during the Gaza war. The disclosures reveal how modern information warfare increasingly relies on deception rather than persuasion.
By Omar Afra
No bullshit, no bananas: Someone got paid to write that Facebook comment.
While the world watched apartment blocks collapse in Gaza and governments argued over ceasefires, an Israeli political consulting firm was quietly operating hundreds of fake social media accounts posing as Americans and Canadians. According to Meta, those accounts coordinated messaging across Facebook and Instagram using fabricated identities. According to OpenAI, the same operation used artificial intelligence to manufacture articles, comments and political content at industrial scale. Investigative reporting by Haaretz, The New York Times and Politico later connected the operation to a reported $2 million contract with Israel’s Ministry of Diaspora Affairs, which has denied directing the campaign of course.
Read that shit again because it should have detonated far louder than it did.
A government-linked influence operation was reportedly spending public money to impersonate ordinary Western citizens while flooding political conversations with synthetic speech during one of the bloodiest conflicts of the century. That is not ordinary public relations. It is an information operation.
The trick-fuck wasn’t simply defending Israel. Governments defend themselves all the time. The trick-fuck was understanding that the easiest way to reduce scrutiny of Israel’s conduct in Gaza was to change the subject entirely. If the conversation becomes about how scary the Muslims are instead of genocide, Sharia instead of starvation, the political terrain shifts without anyone noticing the ground has moved beneath their fucking feet.
Meta’s investigation found hundreds of fake accounts. OpenAI independently documented the use of generative AI to increase the speed and volume of the operation. Reporting found personas presenting themselves as Black Americans, Jewish students, concerned Canadian citizens and ordinary voters while promoting narratives about radical Islam, circulating false claims involving Sharia law and attacking Palestinian advocacy organizations. The operation reportedly targeted members of Congress, journalists and online communities already engaged in debates surrounding Gaza.
That is not bullshit scraped from anonymous Telegram channels. Those findings came from platform investigations, corporate disclosures and reporting by major news organizations.
The obvious temptation is to treat Stoic as the scandal and move on. That would be a mistake. The real story is what Stoic tells us about the industrialization of political influence and how feckless and vulnerable your aunts and uncles are on Facebook.
The reported contract tied to Stoic was worth roughly $2 million. By itself, that is no small potatoes. Inside Israel’s rapidly expanding public diplomacy apparatus, however, it begins to look less like the entire operation than one receipt somebody accidentally left on the table. Reporting over the past two years has documented dramatic increases in Israeli spending on international public diplomacy alongside additional government-supported initiatives aimed at influencing discourse abroad. Subsequent reporting uncovered new campaigns involving influencers, strategic communications firms and targeted outreach across the United States.
Nobody should assume every dollar funded covert social media operations. There are plenty of dirty tricks necessary elsewhere. But nobody should seriously believe a government dramatically expanding its influence infrastructure hired exactly one contractor, built exactly one fake network, got caught once and packed up the entire operation and called it a day. That is not how proper propaganda functions.
The pattern is familiar. Russia’s Internet Research Agency was initially understood through a handful of fake Facebook pages before investigators uncovered a far broader ecosystem spanning multiple platforms, shell companies, and years of operations. Chinese influence campaigns have likewise emerged incrementally, with Meta, Google, Microsoft, and independent researchers identifying new networks over successive investigations rather than all at once. The same has been true of Iranian operations, where individual takedowns gradually revealed a larger collection of contractors, media fronts, and coordinated online assets. The Stoic disclosures should be viewed through that same lens. Meta documented one network. OpenAI identified one AI-assisted operation. Journalists traced one reported contract. Whether those disclosures represent a substantial portion of Israel’s broader influence activities or only the first publicly documented fragment remains an open question. History suggests that sophisticated information operations are rarely exposed in a single revelation; they tend to become visible one investigation, one contractor, and one tranche of documents at a time.
That is why Stoic deserves to be understood as a window rather than an exception. The campaign also exposes something uncomfortable about how digital propaganda works. It does not need to manufacture hatred from scratch. It only needs to find existing prejudice and feed it better material. In conservative regions with growing Muslim populations, such as Texas, the soil was fertile for knee jerk Islamophobia.
Anti-Muslim politics did not begin on October 7. Decades of post-9/11 rhetoric, domestic culture wars, anti-immigration campaigns and algorithmic outrage had already built an enormous audience prepared to consume stories portraying Muslims as existential threats. Stoic, if the public reporting is accurate, appears to have recognized that reality and exploited it. That distinction matters. The operation was not inventing Islamophobia but was attempting to operationalize it.
Researchers have repeatedly found measurable relationships between online anti-Muslim hate speech and real-world harassment and violence. Civil-rights organizations recorded sharp increases in anti-Muslim incidents after October 7. No responsible investigation can attribute those developments to one covert campaign. The evidence does not support that. What the evidence does support is something more than troubling enough: a documented influence operation deliberately injected anti-Muslim narratives into an environment already saturated with fear, anger and grief. Pouring gasoline on a fire is still an act even if someone else struck the first match.
The online rhetoric has unfolded alongside a troubling series of real-world attacks. In recent months, New York police opened a hate-crime investigation after a man allegedly threatened two young Muslim women on a Queens bus while shouting anti-Muslim slurs and brandishing a box cutter. Earlier this year in Brooklyn, prosecutors charged a woman with hate crimes after police said she assaulted three Muslims—including a 12-year-old girl wearing a hijab—in an 11-minute spree. In Canada, Toronto police launched a hate-crime investigation after a man was captured on video allegedly kicking and verbally abusing a hijab-wearing woman on a Durham Region Transit bus, repeatedly referring to her as a “terrorist.” Ottawa police are also investigating an alleged hate-motivated assault in which a woman’s hijab was forcibly removed, while Edmonton police recently recommended hate-crime sentencing enhancements after a woman wearing a hijab was allegedly targeted in a racially and religiously motivated hit-and-run. None of these incidents can be causally linked to any particular online influence campaign, but together they underscore a documented reality: anti-Muslim hostility is not confined to social media feeds. It is increasingly spilling into buses, sidewalks, schools, and public spaces, with visible Muslim women often bearing the brunt of that violence.
That is where this story stops being about hillbillies on Facebook.
Information operations have consequences because people mistake them for society talking to itself. A fake account pretending to be your neighbor is not trying to convince everyone. It is trying to convince enough people that everyone else already agrees. Once authentic users begin repeating the message, the operation no longer needs the fake account. The algorithm takes over and the audience finishes the job. The original authors disappear into the background while everybody else argues over conclusions they quietly helped manufacture. rarely do you see those who post respond to rebuttals.
The bullshit is not only the fake account. The bullshit is convincing millions of people that what they are witnessing is spontaneous public opinion when parts of it have been deliberately engineered.
Whether Stoic ultimately proves to be the largest Israeli influence operation or merely the first one investigators managed to expose remains an open question. The evidence available today cannot answer it. What it can answer is far more significant than many seemed willing to acknowledge. A government-linked operation reportedly spent public resources to fabricate Western citizens, inject anti-Muslim narratives into democratic debate and manipulate conversations surrounding an active war. That should not be remembered as an embarrassing public-relations mishap. It should be remembered as one of the clearest demonstrations yet that modern geopolitical conflict is fought as aggressively through information as through military force.
The next time your feed suddenly fills with dozens of strangers repeating the same emotionally charged story, the first question should not be whether you agree with them. It should be whether they exist at all.
Omar Afra is a writer, cultural producer, and founder of Testset. His work focuses on culture, technology, power, war, and the infrastructure beneath public life. He also builds cedar pergolas. More at testset.media/authors/omar-afra.

Comments