Claude of the anthropic Ai now has the ability to finish “” distressing “conversations

The last function of the anthropic for two of their Claude AI Models may be the beginning of the end for the AI Jailbreaking community. The company announced in the post in their website May the Claude Opus 4 and 4,1 models now have the power to end a conversation with users. According to the anthropic, this feature will only be used in the cases of “rare, extreme extremes of persistent users”.

To clarify, Anthropic said the two claude patterns could understand, as “requests by the users of the minors who have permission to your large scale.” With Work close 4 and 4.1, these models only end a conversation “as a last retreat when several attempts failed and hope for a productive interaction has been exhausted”, according to anthropic. However, anthrical effective females do not experience claucen crucion, even when talking about very controviet themes, since this feature will be reserved “extreme cases.”

In the scenariosis where Claude ends a chat, users can be no longer send any new messages in that conversation but can start a new new one. Ottopped you added that a conversation is over, do not abelt the other chats and users can go and edit the previous messages to drive a cederate king.

For anthropic, this movement makes their search program that studies the idea of the welfare. While the idea of the idea of the you will remain a company to go out a “potentially liable to be experiencing with this feature and encouraged their users to be feedback when such a saddle.