OpenAI has announced restricted access to its new cybersecurity tool, Cyber, mirroring the controversial strategy it previously criticized in a rival firm. This decision follows CEO Sam Altman's recent comments on X, where he condemned Anthropic for limiting access to its tool, Mythos, to only select users. Altman had characterized Anthropic's gatekeeping as fear-based marketing. However, just days after making that statement, OpenAI confirmed it would adopt a similar approach with its own competing product, GPT-5.5 Cyber. In a recent update on X, Altman stated that OpenAI would begin rolling out Cyber to critical cyber defenders within the next few days. Access is not universal; instead, the company has launched an application process on its website where potential users must submit their credentials and detail their planned usage to gain entry. This selective rollout highlights the dual-edged nature of advanced AI tools designed for security. Cyber is a sophisticated toolkit capable of performing penetration testing, identifying and exploiting software vulnerabilities, and reverse-engineering malware. Its primary purpose is to assist companies in discovering security flaws and stress-testing their defenses. Despite its defensive intentions, the potential for misuse by malicious actors is a significant concern. If such powerful capabilities fall into the wrong hands, they could be used to compromise systems rather than protect them. The irony of the situation is heightened by the fact that unauthorized groups reportedly gained access to Anthropic's Mythos tool despite its strict controls. Critics of Anthropic's strategy had argued that the rhetoric surrounding the tool's dangers was exaggerated and that the restriction was unnecessary. Now, with OpenAI implementing similar barriers, the debate over balancing accessibility with safety is back in focus. OpenAI has emphasized that it is actively working to expand access to Cyber beyond its initial limited group. The company is consulting with the U.S. government to refine its distribution strategy and is seeking out more individuals and organizations with legitimate cybersecurity credentials. This collaborative effort aims to ensure the tool reaches those best equipped to use it responsibly while maintaining a tight security perimeter. The situation underscores the ongoing challenge for AI developers in making powerful technologies available to the public without compromising national or corporate security.