Security researchers in an OpenAI bug bounty program hacked OpenAI, accessing its "monorepo" on GitHub, using a cybersecurity version of Opus 4.8 and Opus 5 (Robert McMillan/Wall Street Journal)

Security researchers in OpenAI’s bug bounty program were able to access its internal GitHub monorepo using AI-assisted cybersecurity tools, underscoring how quickly automated offensive capabilities are maturing. For CIOs and technology leaders, the business impact is clear: AI is lowering the barrier to sophisticated intrusion attempts, increasing the risk of intellectual property exposure, source code compromise, and operational disruption even at well-defended organizations. IT organizations should treat AI-augmented attackers as a mainstream threat vector and strengthen repository security, least-privilege access, secrets management, and continuous red-team testing accordingly.