A recent investigation by fact-checkers has uncovered a troubling trend: numerous Facebook pages are posting AI-generated portraits and biographies of supposed Holocaust victims, using fabricated details to drive engagement and potentially earn money. These posts, which often present fake images alongside emotional stories, are being condemned by historians and memory institutions as a cynical exploitation of historical suffering.
The scheme relies on artificial intelligence to create fictitious yet poignant narratives—such as a victim who “loved her violin” or “sang lullabies to her brother”—that are nearly impossible to verify. In some cases, real photographs from the Auschwitz Museum have been altered with AI, while names from Israel’s Yad Vashem archives are paired with made-up birthdates, hometowns, and death accounts.
Fact-checkers identified more than a dozen Facebook pages engaged in this practice, with the primary motivation appearing to be financial gain. By posting high-engagement content that garners likes, comments, and shares, these pages may qualify for Meta’s invitation-only “Performance Bonus” program, which pays creators based on interaction. One estimated calculation suggested that a single post with thousands of likes and comments could generate around $77 in revenue at minimal cost.
The Auschwitz Museum issued a strong condemnation, calling the posts “a profound act of disrespect to the memory of those who suffered and were murdered.” The museum noted that the pages replicate real biographical data from its own posts but pair it with fake, AI-generated images that mislead viewers. It urged the public to verify sources and resist the spread of manipulated historical content.
Some of the pages behind this content appear to be managed by entities with dubious backgrounds. One page, previously run by a U.S. fire marshal’s office, later changed hands. Others claim to be managed by U.S.-registered companies, but no records for these firms could be found in commercial registries. Administrators are reportedly located in countries including Australia, Nigeria, Iran, and India.
The volume of posts—between 10 and 50 per day—suggests automation in both content creation and scheduling. While Meta did not respond to requests for comment, indicators suggest these pages are profiting from the engagement. Meanwhile, the company has phased out its third-party fact-checking program in the U.S. in favor of a community-driven system, but none of the AI-generated Holocaust posts reviewed had been flagged or corrected by users.
As AI-generated content continues to proliferate—with nearly one in four Facebook posts between 2023 and 2024 being AI-produced—the challenge of distinguishing genuine memory from digital fabrication grows ever more pressing.