After publicly criticizing Anthropic for restricting access to the Mythos cybersecurity tool, OpenAI is again following a similar direction with a competing product called Cyber, which is an artificial intelligence tool developed by OpenAI, aimed at the field of cybersecurity.
This move shows contradictions in how AI companies balance technology innovation and safety risks.
On social network X, CEO Sam Altman confirmed OpenAI will deploy GPT-5.5 Cyber to a limited user group, mainly important cybersecurity experts.
Access rights are not widely expanded but require registration and provision of professional information and purposes of use before being considered.
According to the description, Cyber is a powerful toolkit that supports intrusion testing, detection and exploitation of security vulnerabilities, as well as malware analysis.
This tool is designed to help businesses assess system defense capabilities. However, that very power also raises concerns that it could be exploited for cyber attack purposes.
Previously, when Anthropic applied a restrictive strategy to Mythos, Altman criticized this as a "marketing based on fear". Some opinions in the technology world also suggest that Anthropic's warnings are somewhat exaggerated.
However, in reality, groups have emerged seeking to approach Mythos illegally, highlighting the risk of controlling sensitive technology.
In that context, OpenAI's decision shows that even critics must be cautious when facing tools that can have a major impact.
The company said it is coordinating with US authorities and expanding the list of eligible users, in order to both ensure safety and gradually popularize technology.
The competition between AI companies in the field of cybersecurity is entering a sensitive stage, where the boundary between innovation and control becomes more fragile than ever.
