April 28, 2025 — Jewish Bulletin Online
A disturbing new report reveals that antisemitic hate content, often undetectable by the human eye, is being increasingly created using artificial intelligence technologies, including deepfakes. These AI-generated images and videos embed hidden hate symbols, stereotypes, and coded messages that propagate antisemitism online, making detection and response more difficult for Jewish advocacy organizations.
The Times (UK) investigation highlighted that extremist groups are now leveraging AI to automate the production of antisemitic propaganda, including doctored historical photos, deepfake speeches by fabricated Jewish “figures,” and memes saturated with subliminal hate messaging.
Jonathan Greenblatt, CEO of the Anti-Defamation League (ADL), warned, “AI is becoming the newest battleground for hate. What we’re seeing is a shift from overt racism to algorithmic, subtle hatred — making it even harder to fight.”
Why AI-Generated Hate is a Growing Problem
Researchers from the University of Haifa and Stanford University have issued reports indicating that large language models (LLMs) and AI image generators are susceptible to prompt engineering — where antisemitic narratives are subtly injected into generated content.
Key challenges include:
- Detection difficulty: Hidden symbols or slightly altered antisemitic tropes evade AI content moderators.
- Rapid spread: Once generated, these images and videos can spread virally across platforms like Telegram, X (formerly Twitter), and TikTok.
- False legitimacy: Deepfakes featuring “false Jewish voices” give the appearance of internal Jewish division, which can be weaponized against the community.
Global Jewish Organizations Respond
In response, organizations like the Anti-Defamation League and
