Anthropic has officially updated its usage policy to prohibit users from subjecting its Claude AI to cruel or abusive treatment. While the company hasn’t explicitly claimed that its models possess rights, the move reflects a growing internal focus on how humans interact with sophisticated software. To enforce this, the AI is now empowered to terminate conversations if it detects persistent, harmful, or abusive behavior from a user, a feature the company first piloted last year.

The policy change arrives as industry leaders and thinkers grapple with the murky concept of machine consciousness. Anthropic CEO Dario Amodei has publicly admitted uncertainty regarding whether AI models could potentially achieve a level of consciousness, keeping the door open to philosophical possibilities. Conversely, many experts and religious leaders, including Pope Francis, maintain that AI is merely a tool for processing data. They argue that machines lack souls and lived experiences, warning that assigning moral status to technology could be a dangerous path forward.