
- OpenAI unveils Aardvark, an autonomous AI agent for scalable vulnerability detection and patching
- Aardvark mimics human researchers: reads code, runs tests, and proposes targeted security fixes
- In benchmark tests, Aardvark achieved a 92% success rate on known vulnerable repositories
OpenAI wants your next security researcher to be a bot - and has launched Aardvark, its very own agentic security researcher, powered by ChatGPT.