The initiative centers on a new application security service built upon Project Lightwell, an effort backed by a 5 billion dollar commitment from IBM and Red Hat. Unlike traditional code scanning, this service uses OpenAI models to prioritize high-risk areas within application code. The tool operates through IBM Consulting Advantage, which connects client environments to AI models using read-only access and bounded execution to maintain strict governance. Clients can deploy this as a managed service, starting with targeted evaluations before scaling to continuous monitoring.

Mark Hughes, Global Managing Partner for Cybersecurity Services at IBM Consulting, noted that defenders must match the speed of attackers who already use AI to probe and exploit vulnerabilities. Dane Stuckey, CISO at OpenAI, emphasized that the collaboration focuses on accelerating defensive workflows while ensuring the trust and compliance required by governments and corporations. The service is available immediately, with further integrations planned as the partnership evolves.