New study: Dramatic spread of AI-based anti-Semitism on social media

A new and first-of-its-kind report by the CyberWell organization, an independent technological association working to combat anti-Semitism online, reveals alarming data about how Gen AI tools are completely changing the scale and speed of hate content dissemination. The study, which analyzed hundreds of posts published between January 2025 and February 2026, shows how the rising popularity of photo, video and audio models has led to an explosion of sophisticated anti-Semitic content in the digital space.

According to the organization’s data, a verified database of 300 AI-based anti-Semitic content items alone managed to accumulate more than 30 million views and more than 2.8 million interactions on the main social networks.

“The industrialization of hate”: the accessible tools that have become weapons

The research found that hostile actors misused common and accessible tools such as OpenAI’s Sora, Google’s Veo tool, Grok (on the X network) and Suno to mass and industrially produce blatant anti-Semitic content. These tools made it possible to combine visuals, audio and completely made-up narratives together with video clips from real events – something that made the lies and propaganda believable, attractive and much harder to detect.

“Artificial intelligence has fundamentally changed the scale and speed at which anti-Semitism can be produced and spread online,” said CEO and founder of CyberWell, Tal-Or Cohen Montemayor. “Nowadays, it is possible to produce hateful content on an industrial scale and reach millions, while enforcement usually comes only after the content has already been widely distributed.”

The report also states that following widespread misuse of the Sora video model to distribute problematic content, the OpenAI company eventually shut down the tool – a step that illustrates the huge challenge facing AI developers in preventing misuse of the technology.

The main target: children and gaming spaces

One of the most disturbing findings in the report concerns the target audience. About 79% of the anti-Semitic content examined appeared on video-based platforms (TikTok, Instagram and YouTube), with short video formats and viral trends dominating and targeting a young audience.

The research reveals that AI-based music and gaming content, including those associated with popular games such as Roblox and Minecraft, have been repeatedly used to embed anti-Semitic messages in spaces aimed at young people. A large part of the content dealt with the mockery and denial of the Holocaust, and when the children watched this content, the feed also offered them offensive content intended to exploit or drag minors.

Violent anti-Semitism is multiplying

CyberWell began monitoring the AI ​​phenomenon in a collaborative manner following a sharp increase in exposure to this content after the Hamas attack on October 7, 2023. However, the most significant turning point was recorded during the June 2025 conflict between Iran, Israel, and the US, since a sharp and sustained spike in the cross-platform spread of AI-based anti-Semitism was detected.

The engagement patterns measured were extreme:

Sharp gaps in enforcement between social networks

The report presents a complex picture of the supervision and removal mechanisms on the various platforms:

The organization emphasizes that even when enforcement is carried out, it is often delayed and arrives considerably late, and the content remains online long enough to accumulate hundreds of thousands or millions of views before it is removed.

“Platforms must move beyond mere transparent statements and invest in systems capable of identifying harmful narratives on a large scale, including those embedded in audio, visuals and coded language that manage to escape traditional detection mechanisms,” concluded Cohen Montemayor. She called on the technology companies and policy leaders to strengthen the automatic identification mechanisms, invest in high-quality human control and cooperate with external experts to curb the developing wave of digital hate.