IBM has teamed up with OpenAI to launch a new security service that uses advanced AI to find, test, and fix software vulnerabilities before hackers can exploit them.
The new service is part of OpenAI’s “Daybreak Cyber Partner Program.” It connects OpenAI’s specialized security models with IBM’s consulting platform to scan company software. Unlike traditional scanners that often flag harmless errors, this AI can actually analyze code to see if a bug is truly dangerous and exploitable. It operates safely inside a company’s secure environment with read-only access to the code.
This partnership will also boost Project Lightwell, a massive $5 billion security initiative by IBM and Red Hat. Backed by a global team of engineers, Project Lightwell acts as a safety clearinghouse for open-source software. By adding OpenAI’s technology, the project aims to automatically write and test security patches for popular open-source code used by businesses worldwide.
The main goal is to give cyber defenders an advantage, as hackers are already using AI to launch faster, automated attacks. Mark Hughes, head of cybersecurity at IBM Consulting, noted that defenders need these same fast tools but with the strict safety controls that businesses require. OpenAI’s Chief Information Security Officer, Dane Stuckey, added that the tool helps organizations find risks while keeping their data private and compliant.
The new AI scanning service is available to IBM clients starting today.
