Sahar Abdelnabi is a Principal Investigator at the ELLIS Institute Tübingen and an independent research group leader at the Max-Planck Institute for Intelligent Systems and Tübingen AI Center. She leads the COMPASS research group (COoperative Machine intelligence for People-Aligned Safe Systems). She was previously an AI Security researcher at Microsoft and completed her PhD at CISPA Helmholtz Center for Information Security. Her past work was the first to disclose and identify indirect prompt injection attacks, a major and now widely recognized LLM security vulnerability. Her recent work on LLM biases received a best paper award at ACL 2025.

‍