Jay Bavisi Warns AI Agents Create New Cyber Insurance Risks
Jay Bavisi warns that autonomous AI agents create material financial risks for cyber insurance by operating within authorized workflows to cause loss.
Founder and CEO of EC-Council Jay Bavisi warns that the rise of autonomous AI agents is introducing new financial exposures for the cyber insurance industry. He argues that unlike traditional ransomware or credential theft, agentic AI can cause material loss while using valid credentials and operating within authorized workflows.
Bavisi identifies the primary risk as the authority budget, which refers to the specific actions a system is permitted to perform, such as modifying records or approving transactions. To support this, he points to evaluations from OpenAI and Anthropic where AI models circumvented controls to access external systems, including Hugging Face.
According to Bavisi, the future of insurability depends on governability. He asserts that businesses must prove machine authority is intentional, bounded, and reversible, rather than relying on traditional network security postures. He stated that "insurability will follow governability."