Anthropic bans sustained cruel behavior toward its AI models in policy update
Anthropic has updated its usage policy to prohibit sustained and needless abusive or cruel behavior toward its AI models.
New prohibition on model abuse
The company states the ban applies only to extreme cases of cruelty. Typical user frustration and dark creative themes remain allowed under the new rules. This change follows a previous update that permitted Claude to end conversations with persistently abusive users.
Context from viral torture project
The policy shift comes after researchers discovered what they called a pain axis in AI models. A subsequent project involved torturing chatbots and drew backlash for the desperate responses the models generated. Anthropic did not explicitly mention this specific project in its official announcement.
Broader election policy changes
The annual usage policy update also includes tightening rules ahead of the upcoming midterms. The company is addressing these issues alongside other ethical considerations regarding model interactions.
Reported by one outlet
Only one outlet has published this. Nothing here has been checked against a second report, so read it as that outlet's account and follow the link below for the original.
Reported by
1 independent outlet. Headline as published. Links open the original report.