[DigitalToday reporter Chi-gyu Hwang (황치규)] OpenAI has launched an AI-based application security agent, Codex Security, Axios reported on March 6 local time.
The company said Codex Security finds vulnerabilities in code repositories, verifies them and even suggests fixes.
Codex Security is an evolved product based on Aardvark, a security research agent OpenAI has tested since last year with a small number of customers. It works by analysing code repositories, validating suspected vulnerabilities in a sandbox environment, then generating a proof-of-concept exploit to confirm real-world impact and propose remediation steps.
OpenAI is making Codex Security available as a research preview to enterprise, business and education customers. It can be used for free for the first month.
Anthropic also unveiled Claude Code Security last month, and share prices of major security companies fell afterward.
·'AI fear' hits the security industry..."It cannot destroy core businesses"
Chi-gyu Hwang (황치규)
delight@d-today.co.kr