Palo Alto and Cisco launch AI security tools to protect autonomous systems from emerging threats.

Palo Alto Networks has launched Prisma AIRS 3.0 and a new secure browser designed to protect agentic AI systems throughout their lifecycle. The platform offers unified visibility, risk assessment, and runtime security across cloud, SaaS, and endpoints, with features like Agent Artifact Security, AI Red Teaming, and a preview of the AI Agent Gateway. The secure browser enforces content-aware boundaries, blocks malicious prompts, and distinguishes human from AI activity to prevent data leaks. Cisco also introduced AI security enhancements at RSAC 2026, including Zero Trust Access for AI agents, the open-source DefenseClaw tool for vulnerability scanning, and free security tools to support secure AI adoption. Both companies are advancing AI-specific security to address growing risks from autonomous agents, shadow AI, and prompt injection.