On December 11, 2025, OpenAI warned that its upcoming artificial intelligence models could pose a "high" cybersecurity risk as capabilities advance at breakneck speed. The company cautioned that future models might develop working zero-day remote exploits against well-defended systems or assist with complex enterprise and industrial intrusion operations aimed at real-world effects.
To counter these threats, OpenAI is investing in strengthening models for defensive cybersecurity tasks and creating tools that enable defenders to more easily perform workflows such as auditing code and patching vulnerabilities. The approach combines access controls, infrastructure hardening, egress controls and monitoring to mitigate potential misuse.
Later this year, OpenAI will introduce a tiered-access program for qualifying users and customers working on cyberdefense, granting enhanced capabilities to those on the front lines of digital security. This initiative aims to empower the defenders rather than the threats.
OpenAI is also establishing the Frontier Risk Council, an advisory group that will bring experienced cyber defenders and security practitioners into close collaboration with its teams. Initially focused on cybersecurity, the council will expand into other frontier capability domains in the future, guiding safe and responsible AI development.
As AI technology pushes boundaries, the industry’s collective vigilance will be critical to ensure these powerful tools serve as shields rather than weapons. Stay tuned for updates on how OpenAI and its partners are shaping the future of cyberdefense in the AI era.
Reference(s):
cgtn.com
