Anthropic bans abusive behavior toward AI models in policy update

Anthropic has updated its usage policy to prohibit “sustained and needless abusive behaviour” toward its AI models, Inc42 reported. The change introduces formal restrictions on interactions that Anthropic describes as cruel or harassing.

The policy update follows a series of cost-cutting model releases from the company, including recent versions of Claude Haiku and Sonnet, both launched at significantly reduced prices compared to their predecessors. Earlier coverage of these updates highlighted speed and efficiency; now the conversation has shifted to safety and governance.

Anthropic’s move mirrors broader industry trends. As AI models become more human-like in their responses, providers are grappling with how to balance open access with protections against misuse. The question is whether these rules will be enforceable at scale, or if they risk alienating developers who rely on unrestricted model testing.

What comes next will matter most for enterprise users. Companies already shifting from AI pilots to private, storage-backed models may see Anthropic’s policy as a signal that safety and cost control are converging. For now, the update raises an open question: can a platform enforce behavioral standards without limiting the experimentation that drives adoption?

Sources: inc42.com