When Facebook users scroll through their feeds and see portraits of supposed Holocaust victims—complete with names, ages, and tragic fates—they often assume these victims are being commemorated . But in many cases, the images and biographies have been generated by AI and fabricated, according to an investigation by the Russian-language fact-checking site Provereno (“Verified”).
The spread of such posts appears to be fueled by Facebook’s content-monetization feature and has been denounced by the Auschwitz Museum as “a profound act of disrespect to the memory of those who suffered and were murdered” at the Nazi death camp.
Last month, the Russian-language Israeli news site Newsru.Co.Il examined a flood of suspicious Holocaust-related posts, most featuring a portrait of a supposed victim alongside a short biography. Many included emotionally charged details—such as “loved her violin” or “sang lullabies to her brother”—that would be exceedingly difficult, and likely impossible, to verify.
The site found that many images were generated using artificial intelligence, and that the accompanying text was largely fabricated. In some cases, real photographs of Holocaust victims—previously published by the Auschwitz Museum—were altered with AI. Names from Israel’s Yad Vashem archives were paired with fictional birth dates, hometowns, and death narratives.
While Newsru.Co.Il did not name the groups behind the posts, Provereno identified more than a dozen Facebook pages trafficking in fictitious, AI-generated biographies of supposed Holocaust victims.
According to Provereno, the underlying motive appears to be profit.
These pages may be earning money through Meta’s invitation-only “Performance Bonus” program, which rewards creators for engagement on public posts. Provereno estimated that one AI-generated Holocaust post from the page “90's History” with 9,600 likes, 1,300 comments, and nearly 1,000 shares could have generated roughly $77 in revenue—at minimal production cost.
Just days after the Newsru.Co.Il report, the Auschwitz Museum issued a statement condemning the posts for undermining “the integrity of historical truth.”
“What makes this particularly troubling is that their posts copy real content—including names, dates, and biographical facts taken directly from our posts—yet they pair this information with fabricated, AI-generated images that mislead viewers,” the museum said. “We strongly condemn this practice and urge anyone committed to preserving the memory of Auschwitz to verify sources and stand firmly against the spread of manipulated or misleading historical content.”
‘Performance Bonus’ for Fabricated Posts
In its statement, the Auschwitz Museum singled out “90's History” as one of the pages spreading AI-generated Holocaust content. While other pages examined by Provereno published similar posts, they did not appear to be part of a coordinated network. Some had previously been operated by unrelated administrators.
Until spring 2024, “90's History” was reportedly managed by the fire marshal’s office in the U.S. state of Tennessee. Both it and another similar page, “Historical Snapshots,” claim to be managed by U.S.-registered companies—Star Groups LLC and Being Master LLC, respectively. However, Provereno found no records for these companies in U.S. commercial registries. In fact, “Historical Snapshots” appears to be managed from Myanmar.
Despite this, both pages received “verified” status from Facebook’s parent company, Meta—raising questions about the company’s vetting process.
Meta did not respond to Provereno’s request for comment before publication.
Other pages posting similar content are reportedly managed by administrators in countries including Australia, Nigeria, Iran, and India.
While the timing of the first AI-generated Holocaust posts varies, Provereno noted a spike beginning in April and May of 2025. Most of the posts pulled names and images from the Yad Vashem database, accompanied by fabricated stories generated using AI.
The fact-checkers found that these pages published between 10 and 50 posts per day—a volume suggesting automation for both the creation and scheduling of content. Although participation in Meta’s monetization program cannot be confirmed publicly, Provereno cited several indicators suggesting these pages were profiting from the engagement.
All of the Facebook pages examined had been registered or reformatted after Meta launched its Performance Bonus program.
According to one estimate cited by Provereno, nearly one in four Facebook posts published between 2023 and November 2024 was AI-generated. Since the launch of ChatGPT in late 2022, that number has more than quadrupled.
Meanwhile, Meta has phased out its third-party fact-checking program in the U.S., stating it will be replaced by a system similar to the “Community Notes” model on X (formerly Twitter).
That system is already live in the United States. However, Provereno found that none of the AI-generated Holocaust posts it reviewed had been flagged or corrected by community notes.
