Fable 5 also blocks secure code

▤ Summary

Anthropic’s Fable 5 model is drawing criticism because its safety filters appear to block many legitimate cybersecurity and software development tasks. Researchers say the model flags harmless prompts such as code reviews, secure coding requests, and even reading security-related blog posts. The issue also affects defensive security work, including incident response, detection engineering, and digital forensics. Anthropic says the safeguards are intentionally conservative, but the current behavior may make the model difficult to use for professional security and engineering teams.

▦ Linked Companies

Anthropic

$10M to $25M