Oracle and OpenAI launch AI Security Mesh to find vulnerabilities

The announcement comes as companies patch bugs at unprecedented rates. Oracle recently addressed around 3,000 issues in a quarterly update, most of which were identified with AI assistance. The new AI Security Mesh feature provides supporting evidence, such as reproduction steps and test results, to help users rule out false positives and prioritize critical threats. Altman suggested that this shift may require businesses to move from fixing individual bugs to continuous monitoring, potentially impacting security team structures.

OpenAI Head of Cyber Johannes Landgraf noted that frontier AI offers new opportunities to respond to sophisticated threats. However, a significant concern remains that attackers are using AI to exploit vulnerabilities at the same speed they are discovered. Altman acknowledged that patching every bug individually across the supply chain is difficult and that new methods for remediation are needed.