Tiffany Saade is a product manager for AI security at Cisco and co-author of an April 2026 issue brief on data privacy in foundation-model systems published by the Stanford Institute for Human-Centered Artificial Intelligence (Data Privacy and Foundation Models: Can We Have Both? — King & Saade (Stanford HAI Issue Brief, April 2026)).
Background
Saade graduated from Stanford University's Master's in International Policy program in 2025. At the time of the April 2026 HAI brief she was a Product Manager for AI Security at Cisco (Data Privacy and Foundation Models: Can We Have Both? — King & Saade (Stanford HAI Issue Brief, April 2026)), and a June 2026 MIT FutureTech study lists her affiliation as Cisco AI Defense (Prioritization of Risks from Artificial Intelligence: A Delphi Study of 272 International Experts).
Published work
With Jennifer King, Saade wrote "Data Privacy and Foundation Models: Can We Have Both?", a Stanford HAI Policy & Society Issue Brief released in April 2026 (Data Privacy and Foundation Models: Can We Have Both? — King & Saade (Stanford HAI Issue Brief, April 2026)). The brief's argument is that privacy risk in foundation models is structural rather than incidental: it arises separately at each stage of the model lifecycle — mass scraping of personally identifiable information during pre-training, memorization into weights during training, regurgitation in outputs, and intimate disclosure by users through chat interfaces — so that addressing it at one stage leaves the others open. King and Saade argue that traditional privacy frameworks including the GDPR, the California Consumer Privacy Act and HIPAA do not reach these risks.
The brief opens with a scenario in which a fraudster clones a voice from a YouTube clip and uses the synthetic audio to bypass a bank's voice authentication. It treats the data users disclose through chatbot interfaces, which the authors term "intimate data," as an underexplored privacy frontier, and adds an adversarial stage to its lifecycle map in which prompt injection, data poisoning and model inversion attacks circumvent privacy safeguards and expose personal data (Data Privacy and Foundation Models: Can We Have Both? — King & Saade (Stanford HAI Issue Brief, April 2026)). King is the Privacy and Data Policy Fellow at Stanford HAI (Data Privacy and Foundation Models: Can We Have Both? — King & Saade (Stanford HAI Issue Brief, April 2026)).
Saade is also one of 18 co-authors of "A2AS: Agentic AI Runtime Security and Self-Defense" (September 2025, SSRN), a paper included in the 287-paper systematic review underlying Neumann, Sargeant and Singh's FAccT 2026 study of system prompts as governance instruments (Prompt Governance? On Governing Technologies Governed by Natural Language). She is credited among the several hundred authors of Prioritization of Risks from Artificial Intelligence: A Delphi Study of 272 International Experts (June 2026), in which the expert panelists are listed as authors (Prioritization of Risks from Artificial Intelligence: A Delphi Study of 272 International Experts).
Relationships
- related: Data Privacy and Foundation Models: Can We Have Both? — King & Saade (Stanford HAI Issue Brief, April 2026) — the co-authored issue brief.
- related: AI and Privacy — the subject area of her published work.
- related: Stanford HAI — publisher of the issue brief; Stanford is also where she completed her master's degree.
- related: AI Pre-Release Vetting, Procurement-Driven AI Governance — adjacent topical areas.
- related: Jennifer King — co-author of the issue brief
- related: Cisco Systems — employer at the time of the 2026 brief
- related: Prioritization of Risks from Artificial Intelligence: A Delphi Study of 272 International Experts, Prompt Governance? On Governing Technologies Governed by Natural Language, Prompt Injection
Coverage note
Saade's positions outside the privacy brief are not established from available sources; her other listed works are a co-authored security paper and credited participation in an expert survey.