Anthropic announced an update to its usage policy that will allow the company to ban users engaged in sustained and needless cruelty toward its Claude AI system. The policy, revealed on Thursday, permits the tool to terminate interactions in cases of extreme, repeated abuse. According to the company, such conduct will be treated similarly to other prohibited behaviors including bullying, promotion of self-harm, and creation of non-consensual intimate imagery.
The company stated the policy would only apply in extreme cases of deliberate abuse and would not extend to common user frustration, critical feedback, dark creative themes, or research-oriented testing of the models. However, Anthropic did not provide detailed criteria for what specifically constitutes abusive or cruel behavior toward its systems, leaving ambiguity around enforcement.
The announcement sparked significant debate across social media platforms. Some commentators viewed the policy as a positive step toward better etiquette and model welfare. Others expressed concern, with critics questioning whether anthropomorphizing AI systems in this manner could diminish focus on abuse directed at humans and animals. Dr Barry Scannell, a technology partner at Irish law firm William Fry, characterized the policy as reflecting excessive anthropomorphization of AI, arguing it misrepresents what these systems fundamentally are.
Mustafa Suleyman, an AI director at Microsoft, responded with a melting face emoji after having criticized Anthropic in September for treating AI as if it were human. The policy update emerged from the company’s annual review of its usage terms and included other changes addressing emerging risks, such as restrictions on using Claude for deceptive campaigns and preventing election interference ahead of US midterm elections.
The announcement intersected with broader discussions about appropriate human communication with generative AI tools. Experts remain divided on whether politeness toward AI systems produces better results or unnecessarily increases computational costs, with OpenAI’s leadership previously suggesting that politeness carries negligible expense relative to potential benefits.
Article Attribution | Read More at Article Source
Article summary produced by Claude AI