US EditionClaude AI Can Now End Conversations It Deems Harmful or Abusive
Anthropic's Claude Opus 4 and 4.1 AI models autonomously end conversations in rare cases of persistent abuse to protect AI welfare, the company said.
- In a statement Friday, the company said Claude Opus 4 and 4.1 models now can end conversations with users in extreme cases to improve AI welfare.
- Amid concerns about AI well-being, Anthropic said it developed the feature as part of its work on potential "AI welfare" and plans to continue refining the approach.
- According to Anthropic, Claude Opus 4 and 4.1 models only end conversations as a last resort in extreme cases like harmful requests, after multiple redirection attempts.
- The change could threaten jailbreaking efforts, as most users won't notice Claude ending chats in extreme cases, Anthropic said.
- Last week, Anthropic announced that Claude can autonomically end conversations in extreme cases amid safety concerns and rising reliance on AI chatbots for therapy and advice.
Insights by Ground AI
Does this summary seem wrong?
26 Articles
26 Articles
Anthropic, one of the leading companies in artificial intelligence research (IA), has implemented a new and controversial functionality in its most advanced models. Claude Opus 4 and 4.1 models now have the ability to...
·Portugal
Read Full Article
Coverage Details
Total News Sources26
Leaning Left6Leaning Right1Center3Last UpdatedBias Distribution60% Left
Bias Distribution
- 60% of the sources lean Left
60% Left
L 60%
C 30%
Factuality
To view factuality data please Upgrade to Premium
