TLDR: A new global survey, spearheaded by AI expert Maya Ackerman and Israeli President Isaac Herzog, has been launched to systematically identify and address antisemitic biases within leading generative AI platforms. The ‘Voice of the People’ initiative invites AI users worldwide to report instances of bias, misrepresentation, or antisemitism encountered in AI-generated content, aiming to foster more inclusive and accurate AI systems.
A significant global initiative has commenced with the launch of a new survey aimed at uncovering and rectifying antisemitic biases embedded within artificial intelligence models. This pioneering effort is being led by renowned AI expert Maya Ackerman, a professor of Computer Science and Engineering at Santa Clara University, in collaboration with Israeli President Isaac Herzog. The ‘Voice of the People’ initiative announced on Monday that the survey seeks to compile user experiences of antisemitic content generated by AI.
Professor Ackerman is set to oversee what is described as the ‘first global Jewish initiative dedicated to systematically identifying and combating antisemitism across leading generative AI platforms.’ This undertaking builds upon Ackerman’s prior research, which highlighted concerning patterns in how Jewish identity is portrayed in generative models. Her earlier work revealed that prompts such as ‘Jew,’ ‘Jewish person,’ or even ‘Passover’ could elicit offensive responses, not only perpetuating existing stereotypes but also fabricating new and deeply troubling misrepresentations of Jewish people.
Concerns over AI-generated content misrepresenting religious and ethnic minorities, including Jews, without adequate human oversight, have been growing, according to the Voice of the People. Professor Ackerman clarified that ‘These AI systems are not inherently antisemitic, but they reflect the data they’re trained on—data that is overwhelmingly shaped by a Western internet where antisemitic biases remain embedded and knowledge of Jewish culture is severely lacking.’
The overarching goal of this initiative is to foster the development of more inclusive and equitable AI systems through collaborative efforts and an ‘AI-first approach’ to combating online antisemitism. The survey encourages participants to submit any antisemitic responses they have encountered across various AI tools, including popular platforms like ChatGPT, Grok, Midjourney, and DALL·E. Submissions are requested to include screenshots, the original prompts used, and the user’s emotional reactions to the content.
Also Read:
- Israel’s Generative AI Sector Surges: Over 340 Startups and $20 Billion in Funding
- Africa Launches ACTS AI Institute to Spearhead Ethical and African-Centered AI Development
These collected records will serve as crucial data for guiding new research into systemic AI bias, informing best practices for responsible AI development, and potentially contributing to the improvement of training datasets for these models. Shirel Dagan-Levy, CEO of Voice of the People, emphasized the importance of this work, stating, ‘Jewish identity must be recognized and represented with dignity in the technologies shaping our world.’
