At many organizations, AI agents are no longer experimental. While use of this autonomous technology grows, companies are relying on immature controls to rein it in. Three recent reports detail multiple types of security and privacy breaches already occurring due to organizations’ inadequate policies, limited monitoring and failure to appropriately treat AI agents as risks in their systems. This article, the first in a two-part series on real-world security for AI agents, examines the reports’ benchmarks of incident types, and the safeguards and security measures that companies so far are applying to agents. Part two will provide a playbook for CISOs and their colleagues to strengthen security and reduce risks around AI agents. See “From CEO Deepfakes to AI Slop, AI Incident Tracking Ramps Up” (Jul. 30, 2025).