Researchers are advocating for AI agents to be considered untrusted systems due to inherent risks and potential vulnerabilities that require careful management and oversight.
Researchers: AI Agents Should Be Viewed As Untrusted Systems
By Michael Fawn