World's top AI platforms ignore or deny antisemitism in Persian, ADL research reveals

New research from the Anti-Defamation League (ADL) reveals that leading AI platforms, including ChatGPT and others, fail to identify or reject antisemitic content in Persian as effectively as they do in English. The study tested multiple chatbots and found that none consistently flagged or refused antisemitic statements in Persian, highlighting a significant gap in content moderation across languages. The ADL calls this a systemic failure that could allow hate speech to proliferate unchecked in non-English contexts. The research underscores the challenges of scaling safety measures across diverse languages and cultural contexts in AI systems. The findings were released amid growing global concern over AI-generated hate speech and misinformation.

Global Impact

The ADL research has political and social implications, particularly for platforms operating in the Middle East and Persian-speaking regions. It could accelerate calls for AI regulation that mandates equal safety standards across all languages, not just English.