ICML Foundation Chair Moves from Meta to Google DeepMind to Lead Gemini Security Research

According to monitoring by 1M AI News, renowned scholar in the field of differential privacy, Kamalika Chaudhuri, announced on X that she has joined Google DeepMind as a research scientist. She will lead a research team focused on enhancing the security and privacy of Gemini. She stated that the capabilities of large language models have become sufficiently powerful, and security has become a barrier to fully realizing their potential as agents. She hopes to ‘help Gemini become the safest cutting-edge model.’ Chaudhuri previously served as a research scientist at Meta AI Research (FAIR) and is also a part-time professor at the University of California, San Diego (UCSD). She will assume the role of chair of the ICML Foundation, a leading machine learning conference, starting in July 2025. She is one of the pioneers in applying differential privacy—a mathematical framework that protects individual data from being reverse-engineered—to machine learning training, and in recent years, her research focus has shifted to the privacy and security challenges of AI agents.