New report warns: Terrorists weaponize AI to spread antisemitic propaganda, recruit new members

Artificial Intelligence (AI) is being used on an unprecedented scale by terrorists and others to spread hatred against Jews and Israel, according to a new report published by the American Security Fund (ASF).

“There is a systematic effort by bad actors to spoil AI training data, intentionally making it more antisemitic and anti-Israel,” wrote Julia Senkfor, the report's author. “The ultimate result is AI models that fail to reflect human decency and avoid hate speech.”

Terrorists and other malicious groups are manipulating websites such as Wikipedia, which AI developers heavily rely on to train Large Language Models (LLMs), creating anti-Jewish bias in AI systems. According to Senkfor's research, the distorted information is then taken at face value by the people who encounter it online.

“Recent studies on human perceptions of AI suggest humans are staggeringly overconfident in AI-generated content,” said Senkfor. “An Elon University study found that AI bots are more persuasive than humans in changing human minds on divisive topics, partly because almost half of AI users (49%) believe that AI models are at least somewhat smarter than themselves. Researchers in Germany found that people attribute similar levels of credibility to AI-generated and human-authored content. And most concerningly, a psychological study found that biases introduced by AI can persist in human thinking.”

The Anti-Defamation League found that AI was a significant component in contributing to the more than 300% surge in antisemitism in the United States following the Oct. 7 massacre in southern Israel, according to the ASF report.

The report found that terrorist networks – including Hamas, al-Qaeda, the Islamic State (ISIS) and Hezbollah – are evading online content moderation through the use of AI. They have been generating AI-produced “target identification packages” containing photos of Jewish centers in major cities such as New York, Chicago, Miami, and Detroit, and using AI to recruit new members, including the use of voice-cloning tools.

“In 2023, the Islamic State published a tech support guide to securely using AI tools. Terrorist groups have posted ‘help wanted’ ads to recruit AI software developers, video producers, and open-source experts. ISIS and Al-Qaeda have adopted AI voice cloning software to produce western-attuned news programs in Americanized English. And extremist groups have begun to employ ‘interactive recruitment,’ wherein AI-powered chatbots interact with potential recruits by providing them with tailored information based on their interests and beliefs, thereby making the groups’ messages seem more relevant to them,” the report stated.

Radical groups on the political right, the report revealed, are using AI to mass-produce Jew-hating imagery, so-called “GAI-Hate Memes,” combining antisemitic imagery with satire.

“On platforms like 4chan, users actively share instructions to employ AI image generation tools to create antisemitic depictions, often drawing on traditional antisemitic tropes. And in far-right forums, users discuss creating novel AI models, manipulating existing AI systems, and bypassing mainstream AI safeguards to generate hateful, harmful content…Their rapid mobilization suggests they are transforming a technological innovation into a weapon for mass production and distribution of antisemitic propaganda, fundamentally altering the scale and speed at which hate is being created and disseminated,” Senkfor wrote.

She concluded with a warning about the growing weaponization of manipulated content.

“Extremists are injecting manipulated content into open-editing platforms, especially Wikipedia, to spread and contaminate AI training datasets. Deliberate distortions about Jews and Israel thus evolve into perceived truths. Meanwhile, extremists demonstrate both their willingness and ability to leverage AI. Addressing AI-enabled antisemitism will require understanding it not as an isolated technical problem, but as a complex socio-technical issue reflecting deep historical patterns of prejudice and, at the same time, novel vectors to amplify and spread it at enormous scale and speed.”