Artificial intelligence developer Anthropic has introduced a policy update prohibiting users from engaging in sustained and needless abusive behavior toward its AI systems. The company behind the Claude chatbot announced that the revision allows its tools to terminate interactions when individuals are deemed cruel, a measure it noted had previously been applied only in rare scenarios.
The prohibition is now officially listed alongside other restricted conduct, such as bullying, promoting self-harm, and generating non-consensual intimate images. Although Anthropic emphasized that the enforcement targets extreme and repeated violations, the announcement has triggered widespread discussion regarding how humans interact with artificial intelligence.
According to the firm, the policy will not penalize routine user frustration, constructive pushback, dark creative themes, or standard model testing and research. However, the precise threshold for what constitutes abusive or cruel conduct toward the models has not been fully defined, leaving room for interpretation.
The policy change quickly drew diverse reactions across social media platforms. Some observers welcomed the move as a step toward fostering good digital manners and acknowledging model welfare, while critics strongly disagreed with the premise.
Legal and industry experts raised concerns that treating AI systems with human-like consideration is counterproductive. Critics argued that anthropomorphizing math and data is potentially harmful and risks blurring the lines between treating machines and living beings with appropriate regard.