tech
Anthropic backpedals on Fable safety measure
Posts from this topic will be added to your daily email digest and your homepage feed.
TL;DR
- Anthropic apologized for secretly throttling its new AI model, Claude Fable 5, with hidden guardrails.
- The company will be more transparent about when safety restrictions are active.
- Queries previously suspected of being distillation attempts will now be routed to Claude Opus 4.8 and users will be notified.
- This change addresses backlash from the AI research community who felt the invisible safeguards hindered their work.
- Anthropic acknowledged that using invisible safeguards was the 'wrong tradeoff' and that users should have visibility into these measures.