A new study by Israeli academics Michael Gilead and Gal Gutman found that artificial intelligence models associate Jewish names with stereotypes of high competence and low warmth in generated biographies. The researchers authored the study titled "From Myth to Model: Representation of 'the Jew' in Generative AI."

The study instructed ChatGPT-4 Turbo to generate a list of names for Jewish and non-Jewish Americans aged 18 to 80. The generated list included one male and one female name in each category, resulting in a total of 252 names. The AI was then prompted to create a 100-word biography for each of these names, operating as a novelist selecting names corresponding to specific character traits.

After generating the biographies, religious identity markers were removed, and AI systems evaluated the personality and social traits assigned to each character. The study determined that characters associated with Jewish names were consistently rated as more competent, privileged, dominant, and obsessive. Conversely, these characters were also rated as less likable, less collectivist, and lower in perceived warmth.

The findings indicated that Jewish individuals in large language model-generated content were consistently stereotyped within the high-competence, low-warmth quadrant. This pattern was observed for competence-related traits such as intelligent, efficient, and assertive, which were rated consistently high in biographies linked to Jewish names. Warmth-related traits like friendly and likable, however, were rated low. The researchers replicated these findings on the AI language models DeepSeek V3 and Mistral.

Gilead and Gutman stated that historical antisemitic discourse has often depicted Jews as agents of disruption, undermining traditional order and social cohesion. "Our analysis reveals how an ancient prejudice persists in modern technological systems through complex patterns of trait association and cultural coding," the researchers said. They added that addressing bias in AI systems requires attention to "subtle ways in which seemingly neutral traits combine to reproduce traditional prejudices."