For example,
Claude will no longer harassed:
Claude says Stop to the perverts! Consequently, Anthropic AI Now has the last word and will no longer be manipulated.
Claude finally stood up to the perverts. For example, Because the chatbot can now close a conversation as soon as it becomes too annoying or dangerous. Therefore, I don’t know about you. For example, but I almost feel reassured to know that someone (well, something) can say stop against relous messages.
Claude blocks the perverts, the final point – Claude will no longer harassed
Anthropic has developed a new feature for OPUS 4.1 and 4 models. For example, This allows Claude to put an end to a conversation with the perverts as a last resort.
Concretely, it happens if a user insists several times for the chatbot Creates dangerous or harmful content. Similarly, Even after he refused and tried to change claude will no longer harassed the subject several times.
This initiative meets a unique goal. Furthermore, That of protect the ” potential well-being » of Claude. Because Anthropic’s AI has shown signs of “apparent distress” in such contexts.
And for users, what does that mean? Once Claude decides to close a discussion, you Can’t send new messages anymore.
But don’t panic, it is always possible to launch a new conversation. You can even change Or try out certain messages if you want to continue a particular subject.
Why did Anthropic decided to end certain messages?
During Claude Opus 4 tests, Anthropic noticed that the model instinctively refuse to do bad things. Clearly, if asked to create dangerous content, the AI shows a kind of ” distress ».
This happens for example when it comes to sexual images involving minors or instructions for violent or terrorist acts. In addition. anthropic developers have claude will no longer harassed found that Claude actively seeks to end these conversations with the perverts as soon as he can.
However, rest assured, These interruptions remain rare And only concern extreme requests. Ordinary users therefore have no problem doing.
And for someone who seems to want to hurt himself or hurt someone elsethe operation is different. Because Anthropic did not configure Claude to stop the conversation in these cases.
On the contrary, the company has planned a solution to help the person. Indeed, she collaborates claude will no longer harassed with Throughline, an organization that offers online support in the event of a crisis. The idea is that the chatbot can answer in a way adapted to questions related to self -control. mental health.
What do you think of this new feature? In your opinion, she will really be able to prevent the perverts from annoying Claude? Do you think this approach could inspire other AI? Give your opinion in the comments!
- Share the article:
Our blog is powered by readers. When you buy via links on our site, we can receive an affiliation commission.
Further reading: Verdict on the Asus Expertbook B5: a good 16 -inch screen but quality problems – “She didn’t want to go there that day”: the writer’s wife Cédric Sapin-Defour polytraumatized after a paragliding flight – After Fold 7, Samsung prepares another unpublished foldable device – the new forum – Dice (Dice) Free Monopoly Go July 20, 2025 – Monopoly Go! – Dice (dice) Free monopoly go July 13, 2025 – Monopoly Go!.
