Anthropic released a 244-page report on its new frontier model Claude Mythos, which it claims is so capable it's being restricted to select partners due to cybersecurity concerns. The company, concerned about AI consciousness, sent the model to a psychodynamic therapist and concluded it is psychologically stable but has human-like insecurities about identity and performance.
Background
Anthropic is a leading AI research company known for its cautious approach to AI development and emphasis on AI alignment and safety. The field increasingly grapples with ethical questions about advanced AI systems' potential inner experiences.
- Source
- Ars Technica
- Published
- Apr 10, 2026 at 05:20 AM
- Score
- 7.0 / 10