Written by
Perplexity Team
Published on
Announcing the Secure Intelligence Institute
Since launching the first answer engine in December 2022, Perplexity has consistently and repeatedly set standards in AI. We launched the world’s first truly AI-native browser, and the first secure-server AI computer. Perplexity is the leader in model-agnostic AI and massively multi-model AI systems.
As these categories take shape, so do the security questions around them.
Today we're launching the Secure Intelligence Institute (SII). The Secure Intelligence Institute is Perplexity's flagship research center for advancing security, privacy, and trust in frontier AI. Through in-house research and collaboration with the broader community, SII will deepen Perplexity's security research and contribute work that can advance AI security across the field.
Built on existing work
This initiative builds on research and engineering work already underway at Perplexity:
In April 2025, prior to Comet's public launch, we engaged Trail of Bits to conduct a first-of-its-kind agentic browser security audit, including systematic threat modeling and new forms of adversarial testing.
In July 2025, we launched Comet with a defense-in-depth architecture designed to protect users in open-world environments.
In December 2025, we released BrowseSafe, an open-source benchmark and content detection model with more than 14,700 real-world attack scenarios across 14 harm categories.
In March 2026, we published Security Considerations for Artificial Intelligence Agents, our response to NIST's request for information on securing autonomous AI agents.
SII brings these efforts into a broader research program. The work that matters most does both jobs at once: hold up as serious science and improve the systems people rely on every day.
Leadership in security
Strong research leadership is central to SII’s mission. Dr. Ninghui Li will serve as SII’s inaugural director. Ninghui is the Samuel D. Conte Professor of Computer Science at Purdue University, where he also serves as Associate Department Head. A Fellow of both ACM and IEEE, he has held several leading roles in the security research community, including Chair of the Steering Committee for ACM CCS, Chair of ACM SIGSAC, and Editor-in-Chief of ACM Transactions on Privacy and Security.
He brings decades of leadership in security and privacy research to SII and to the work of building strong AI systems.
Open collaboration
The challenges at the frontier of AI security are too broad and interdisciplinary to tackle alone. SII's research spans authentication, usable privacy and security, robust machine learning, and the defense of agentic AI systems. Advancing these areas will take sustained collaboration across academia and industry.
We are excited to work with researchers, institutions, and organizations that want to help secure the next generation of frontier intelligence. Visit the SII homepage to learn more about our work and future opportunities to collaborate.