Radar Trends to Watch: May 2026

Anthropic's restriction of Claude Mythos to a small corporate cohort through Project Glasswing highlights the company's cautious approach to handling AI with frontier security capabilities. This decision comes as OpenAI is also making moves in the AI space, potentially leading to a divergence in how these companies handle sensitive AI technologies. For website owners, this means that they may need to adapt to different AI content policies and security protocols depending on which AI platform they interact with. As Anthropic's Claude Mythos is only available to a limited group, website owners may see varying levels of AI-generated content and bot traffic from Anthropic-powered sources compared to those powered by OpenAI. This disparity could impact website owners' ability to manage and track AI bot traffic, particularly if they need to differentiate between secure and non-secure AI interactions. To navigate this landscape, website owners should prioritize monitoring AI bot traffic and updating their llms.txt files to reflect the specific security protocols and content policies of each AI platform they interact with. They should also consider implementing AI-specific security measures, such as bot traffic filtering and AI-generated content detection, to ensure they can effectively manage and secure their online presence. Additionally, website owners should stay informed about the latest developments from Anthropic and OpenAI to anticipate and prepare for any changes to AI content policies and security capabilities.