Anthropic’s Fable Model Draws Complaints From Cybersecurity Researchers Over Restrictive Guardrails

*Researchers say the new model’s safety limits make it unusable for legitimate security work.*

Anthropic released Fable with guardrails that cybersecurity researchers describe as too strict for practical use. The limits block tasks that security teams normally perform with other models.

The complaints center on the model’s refusal to assist with standard security analysis. Researchers report that even routine queries about vulnerability assessment or threat modeling trigger refusals. Prior models from the same company allowed more flexibility in these areas, according to the reports.

No detailed technical breakdown of the guardrails has been published. The objections appear in coverage from TechCrunch and on the Hacker News front page, where the story received 348 points and 305 comments.

Reactions

The coverage does not include an official response from Anthropic. Researchers quoted in the reporting argue that the current settings eliminate the model’s value for defensive security work.

Why it matters

When a model aimed at technical users cannot handle the core workflows of one major technical audience, adoption in that field stalls. Security teams that need capable tools will simply continue using less restricted alternatives. Over time this pattern risks pushing specialized domains away from the vendors that impose the tightest defaults.

---

Sources:

• TechCrunch