The Hoover Institution's visibility in policy research and global affairs makes our institution an attractive target for state sponsored activity and AI amplified impersonation threats. This session examines how public information is collected, interpreted, and weaponized against academic communities. Attendees will see live OSINT findings that demonstrate how much can be learned about scholars, leaders, and staff from open sources alone, and how this information enables targeted phishing, impersonation, and reputational attacks.
The session also explores how modern AI models build persona level profiles from public data and how adversaries use these models to generate convincing audio and video impersonations. Real world deepfake case studies illustrate the financial and reputational impact on major institutions. I will demonstrate a controlled red team demonstration and will show how an audio deepfake was created using publicly available material, how video generation platforms detect and block protected individuals, and how quickly attackers can pivot to an unprotected target. A short deepfake clip will be shown to highlight what is possible and why this risk matters for Hoover’s credibility.
Key takeaways include understanding OSINT exposure, recognizing AI generated impersonations in academic and policy environments, safeguarding personal and institutional digital footprints, and communicating cyber risk in terms that resonate with leadership, scholars, and faculty.
