Apple published recordings and a research recap from its 2026 Workshop on Privacy-Preserving Machine Learning & AI, showcasing advancements in secure AI development. The two-day event gathered Apple researchers and external experts to discuss privacy protection in machine learning systems, with a focus on federated learning and foundation models.

Apple's presentations covered critical areas including Private Learning and Statistics, Foundation Models and Privacy, and security attacks. Four key talks were featured, including 'Crypto for DP and DP for Crypto' by Research Scientist Kunal Talwar, alongside sessions from University of Toronto and Georgetown University researchers. The workshop highlighted 24 published works, with three papers originating from Apple's own research teams.

The research explores methods to develop AI while protecting user data, addressing challenges like data memorization in foundation models and privacy accounting techniques. By open-sourcing this research, Apple aims to advance industry-wide privacy standards and encourage collaboration on secure machine learning technologies. The recordings provide insights into cutting-edge privacy-preserving techniques shaping the future of AI development.