Securing AI agents by intent rather than identity
Overview
O'Reilly Radar argues that autonomous AI agents break traditional web security because their browser activity looks the same as a human user's, and that cryptographic signatures prove only which agent is acting, not what it intends to do.
The article recommends treating agent access policy as a commercial decision that security teams then implement. Its proposed defenses are short-lived machine credentials, cryptographic verification through Web Bot Auth, intent detection at the browser layer, and protections against prompt injection.
The article's recommendations are presented as its own position; the report does not describe any independent testing or adoption of these measures.
Written by AI from the articles below · updated Oct 9, 7:50 AM ET
Check the sources:
Article timeline
The articles in this story. Times are ET.
- O'Reilly RadarIntent, not identity: securing AI agents against nonhuman traffic
AIAutonomous AI agents break traditional security models because their browser-based activity looks identical to a human user's, and signatures prove identity but not intent. The article says organizations should treat agent policy as a commercial question with a security implementation, and recommends short-lived machine credentials, cryptographic verification via Web Bot Auth, browser-layer intent detection, and defenses against prompt injection.
Heat trend
Not enough continuous observations to show a trend yet.