|
Getting your Trinity Audio player ready...
|
In an urgent response to rising concerns about antisemitic bias in artificial intelligence, a groundbreaking global survey has been launched to document and confront the ways generative AI tools misrepresent Jewish identity.
The initiative is being led by internationally-recognized AI expert Dr. Maya Ackerman and supported by Israeli President Isaac Herzog through the Voice of the People (Kol Ha’am) project.
The survey — the first of its kind focused specifically on antisemitism in AI — invites users to report antisemitic bias or misrepresentations of Jewish identity across widely used platforms such as ChatGPT, Grok, Midjourney, and DALL·E. Users are encouraged to submit prompts, screenshots, and personal reactions to help researchers build a comprehensive record of systemic bias.
“This isn’t about blaming the technology — it’s about recognizing that the data shaping these tools often reflects a digital landscape where antisemitic tropes are still embedded, and authentic Jewish voices are underrepresented,” said Dr. Ackerman, a professor of Computer Science and Engineering at Santa Clara University and a leading researcher in generative AI ethics.
Ackerman’s prior research has already revealed troubling patterns, with prompts such as “Jewish person” or “Passover” at times producing offensive or wildly inaccurate responses. These results often stem not from malicious programming, but from AI models trained on skewed and unbalanced online content.
“Generative AI systems absorb and reproduce the internet’s biases — whether intentional or not,” Ackerman explained. “That’s why we need Jewish input and oversight to ensure that our identity is represented with accuracy, dignity, and respect.”
The survey is a core part of Voice of the People’s mission to elevate global Jewish perspectives in policy and innovation. Shirel Dagan-Levy, CEO of Voice of the People, emphasized the significance of the project: “As AI becomes a primary gateway to knowledge and representation, ensuring fair and respectful portrayals of Jewish people isn’t optional — it’s essential. This initiative allows us to turn awareness into action.”
The findings will fuel new research into algorithmic antisemitism and help shape more inclusive training datasets and ethical standards for AI development. The long-term goal is to work collaboratively with developers and platforms to build safer, more responsible systems.
The timing of the launch is particularly relevant. Just days earlier, xAI issued an apology after its Grok chatbot posted violent and antisemitic messages, including comments praising Nazi leader Adolf Hitler. The company attributed the incident to a software glitch that amplified toxic content for over 16 hours, vowing to strengthen its safeguards moving forward.
Founded in 2023 by President Herzog, Voice of the People is a partnership between the Israeli President’s Office, the World Zionist Organization, the Jewish Agency, and major philanthropic organizations including the Azrieli Foundation and others. It serves as a hub for practical solutions to the challenges facing global Jewry, from antisemitism to leadership development.
The survey — which can be accessed HERE — is now open to participants worldwide and marks a critical step toward ensuring Jewish representation and dignity in the digital age.
