OpenAI launches Aardvark AI agent built to protect critical code from cyber threats

OpenAI unveiled Aardvark, an autonomous AI agent powered by GPT-5 that continuously scans code repositories to detect vulnerabilities before hackers exploit them. Unlike traditional security tools, Aardvark reads and interprets code like a human researcher, testing potential flaws in secure environments and automatically generating patches for developer review. In benchmark testing, it identified 92% of known vulnerabilities and has already uncovered issues across OpenAI's systems and open-source projects, earning ten CVE identifiers.