Microsoft and Ben-Gurion University of the Negev announce collaboration on research on security and AI
Microsoft and Ben-Gurion University of the Negev are launching a research project in the fields of AI and security. Specifically, the research aims to identify privacy leaks in language models by studying how the timing of processes can reveal hidden or encrypted information. This collaboration seeks to augment insight and learnings from the extensive efforts Microsoft already undertakes on AI and security through research projects with top academic partners. This partnership was established due the expertise contained at the university's Department of Software and Information Systems Engineering in the field of securing large language models (LLMs).
This partnership aligns with Microsoft’s Secure Future Initiative (SFI), launched in November 2023, which reflects Microsoft’s multi-year commitment to prioritizing security in response to the growing scale and risk of cyberattacks. SFI focuses on three core principles: Secure by Design, Security by Default, and Secure Operations. SFI includes engineering efforts across six primary domains: identity protection, production system isolation, network and engineering system security, threat monitoring, and accelerating post-incident response and recovery.
The Negev region, and particularly Be’er Sheva, has become a thriving hub for cybersecurity and AI innovation, drawing on the expertise of a diverse community of researchers and developers from both the city and surrounding area. The partnership will contribute to advancing the cybersecurity and AI sectors locally and beyond.
Dr. Tomer Simon, Chief Scientist at Microsoft Israel R&D, stated: “We are in the midst of a significant technological revolution in AI, and Microsoft is committed to shaping a safe and responsible future for this technology. This collaboration underscores our continued commitment to driving technological innovation, generating impactful insights for diverse industries worldwide in the fields of security and AI, and fostering further collaborations with Israel’s academy.”
Dr. Yisroel Mirsky, Head of the Offensive AI Lab and a member of the Department of Software and Information Systems Engineering at Ben-Gurion University, brings extensive experience in vulnerability assessments. In the past, he has worked on finding solutions for various challenges, ranging from countering fraud to preventing cyberattacks on emergency numbers in the U.S. “If you think of AI conversations as mail, currently hackers can see the shape of the envelopes. Then, they can use artificial intelligence to interpret the content of the correspondence with a success rate of about 50-60%, and that’s exactly what we intend to fight against," explains Dr. Mirsky.
Dr. Eitam Shitrit, Head of Data Science and Responsible AI in the Industry AI group at Microsoft Israel R&D, noted: “Our investments in Responsible AI are substantial, and go beyond fair and transparent models that prevent discrimination and biases. They also include safeguarding user privacy and ensuring safe usage. This collaboration will allow us to further strengthen these aspects and remain one step ahead of potential attackers.”
