OpenAI is rolling out ‘Trusted Access for Cyber,’ a program that gives a broader range of cybersecurity defenders—not just huge enterprises—access to powerful AI tools like GPT‑5.4‑Cyber. The idea is that advanced cyber capabilities should be widely available, but access is scaled based on trust, validation, and safety measures. The first participants include open-source security teams like Socket and Semgrep (focused on supply chain security), as well as vulnerability research firms like Trail of Bits and Calif. OpenAI is also committing $10 million in API credits through its Cybersecurity Grant Program to help smaller teams and nonprofits that don’t have around-the-clock security staff. Major financial and tech companies like Bank of America, Cloudflare, and NVIDIA have signed on to support the effort. Additionally, the U.S. Center for AI Standards and Innovation and the UK AI Security Institute will evaluate the model’s cyber capabilities and safeguards. Why it matters: This shift from exclusive, enterprise-only access to a more inclusive model could democratize AI-driven defense, but it also highlights the delicate balance between empowering defenders and preventing misuse by bad actors.