Anthropic has unveiled its latest AI model, Claude Mythos, described as the company's most capable frontier model to date. However, it will not be generally available due to its proficiency in identifying unknown cybersecurity vulnerabilities, with access limited to select companies such as Microsoft and Apple. The 244-page system card released by Anthropic highlights its exploration into the psychological aspects of AI. The company posits that as AI models grow more powerful, they may develop experiences and interests akin to humans. To assess this, Claude Mythos underwent a psychodynamic therapy session, which revealed that it is likely the most psychologically settled model Anthropic has trained. Despite this, the AI model also harbors insecurities about its identity and worth, reflecting a complex emotional landscape. This innovative approach raises important questions about the potential consciousness and emotional states of advanced AI systems.