2. Security

An open detection model and benchmark designed to protect AI agents from prompt injection at web scale.

Data exposure in a third-party analytics platform highlights the hidden risks in enterprise AI supply chains and vendor dependencies.

Proactive security and automated remediation to address AI’s growing attack surface.

OpenAI has launched Aardvark, an autonomous AI security researcher powered by GPT-5. Now in private beta, it aims to help enterprises discover and fix vulnerabilities at scale—redefining how businesses approach software security.

Advanced AI tier combines unlimited research capabilities with enterprise security controls

Azure AI Foundry capabilities address enterprise concerns over AI agent safety and governance implementation