[Analysis: Anthropic and OpenAI Expose Security Vulnerabilities, Raising Concerns Over AI Model Safety] Odaily Planet Daily reports that Anthropic and OpenAI have both experienced security incidents, sparking market concerns about the safety of AI models themselves. Anthropic is currently investigating unauthorized user access to its Claude Mythos model. Almost simultaneously, OpenAI was reported to have inadvertently exposed several unreleased models in its Codex application. Analysts believe that such incidents highlight the significant security challenges faced by AI model providers, even those emphasizing cybersecurity capabilities. As AI is increasingly used to defend against cyberattacks, platform security and access control issues have also become critical risk points. Industry opinions suggest that these vulnerability incidents underscore the need for scrutiny of AI companies' security governance capabilities and reflect the gaps in the security systems of rapidly developing AI technologies. (The Information)