John Wilson, Senior Fellow of Threat Research at Fortra, explains in Dark Reading how attackers can socially engineer AI agents through prompt injection, potentially manipulating systems authorized to act on an organization’s behalf. He recommends treating AI agents and third-party software as identities, with clear visibility into their permissions, credentials, triggers, and external connections. His guidance highlights why organizations need strong guardrails and regular permission reviews as AI agents gain greater access to business systems.
"Maintain a central inventory of every system that can act on the company's behalf, including what data and systems it can access, what actions it can perform, what credentials it holds, and who is responsible for it," — John Wilson, Dark Reading
