Workshops Archive

Privacy Foundations for Multi-Institutional Scientific Artificial Intelligence


Workshop: 5th International Workshop on Cyber Security in High Performance Computing

Authors: Olivera Kotevska (Oak Ridge National Laboratory), Sumit Jha (University of Florida), Aurélien Bellet (INRIA), Rui Hu (University of Nevada Reno), Nathaniel Bastian (Johns Hopkins University), Rafael Ferreira da Silva (Oak Ridge National Laboratory), and Ravi Madduri and Kibaek Kim (Argonne National Laboratory)

Abstract: Scientific AI, from foundation models to federated analysis pipelines, is becoming shared infrastructure across national laboratories, universities, hospitals, and industry. Its privacy risks often concern an institution’s participation, strategy, or capability rather than single records. Differential privacy, federated learning, secure computation, trusted execution, and provenance each protect parts of the stack, but their guarantees rarely compose across mixed-trust institutions, access tiers, and autonomous agents. This perspective recasts privacy as an assurance problem defined by six elements: protected asset, observer, channel, permitted disclosure, guarantee, and evidence. A lifecycle claim register for a composite cross-institutional scenario reveals two gaps specific to leadership-class facilities: scheduler metadata exposes institutional participation, and instrument-attached control loops leak research strategy through timing and accelerator contention. Six priorities emerge: institution-level guarantees, agent-communication privacy, cross-tier information flow, privacy-compatible reproducibility, leadership-scale accounting, and instrument side channels. We offer a common form for stating, comparing, and auditing otherwise fragmented privacy claims.


Back to 5th International Workshop on Cyber Security in High Performance Computing Archive Listing

Back to Full Workshop Archive Listing