Anthropic has launched Claude Fable 5, its first 'Mythos-class' AI model that surpasses previous Opus models in capabilities, but with strict safeguards preventing it from discussing sensitive topics like cybersecurity, biology, and chemistry. The model redirects such queries to the older Claude Opus 4.8 and issues warnings to users, with the company acknowledging this may cause some frustration due to false positives in under 5% of cases. This cautious approach aims to prevent potential misuse by malicious actors while still making advanced AI capabilities available to the public.
Background
Anthropic is an AI safety and research company that develops large language models, known for its focus on building safe and controllable AI systems. The company has previously expressed concerns about potential misuse of advanced AI capabilities in sensitive domains.
- Source
- Ars Technica
- Published
- Jun 10, 2026 at 03:20 AM
- Score
- 7.0 / 10