Cisco Publishes State of AI Security 2026

Cisco today publishes the State of AI Security 2026, a flagship report surveying AI threat intelligence, policy, and research and covering developments through late 2025. It finds an adoption–readiness gap—83% of organizations planned agentic AI versus 29% feeling secure—and documents prompt-injection, jailbreaks, AI supply-chain vulnerabilities, MCP agent risks, plus Cisco's open-source scanners and fuzzers. The report calls for stronger governance and tooling.