Agentic Browsers and the Same-Origin Policy

Security Lab faculty members Franzi Roesner and David Kohlbrenner recently spoke with UW News about their work studying the security of emerging agentic browsers, specifically, how these browsers can allow malicious websites to leverage the agent to violate the same-origin policy. Read the article here: “Some agentic AI browsers come with major cybersecurity risks, UW study finds“. More details about the project are available here: https://agent-security.cs.washington.edu/agentic_browsers_sop.html