United States: Artificial intelligence company Anthropic has updated its usage policy to prohibit sustained and unnecessary abusive behaviour towards its Claude AI system, allowing the chatbot to end conversations involving persistent harmful or abusive interactions. The policy update comes amid wider debate over whether AI systems could possess consciousness or experience suffering.
What does the updated policy allow Claude to do?
Anthropic’s revised policy identifies Claude’s ability to terminate conversations as the primary mechanism for enforcing the restriction. The company previously announced that the chatbot could end interactions in rare, extreme cases involving persistent harmful or abusive user behaviour.
The updated policy does not explicitly refer to “model welfare”, a concept concerning whether AI systems should receive protections similar to those considered for living beings.
Why has the policy raised questions about AI consciousness?
The change comes amid ongoing discussion within the technology industry about whether advanced AI systems could become conscious. Anthropic chief executive Dario Amodei said in February, according to The New York Times, that the question remained unresolved and that the possibility could not be ruled out.
Anthropic did not immediately respond to a request for comment, according to AFP.
What are experts saying about machine consciousness?
Jackson Stakeman, a general manager at Atlanta-based AI services provider Sparq, told AFP that debates over AI consciousness may not be the most useful way to assess the technology. He argued that AI systems reflect human inputs and behaviour at scale, providing a basis for considering how they should be used and governed.
His comments highlight an alternative approach to AI policy that focuses on system behaviour and its consequences rather than determining whether a machine is conscious.
Why do some technology leaders reject AI rights?
Microsoft AI chief Mustafa Suleyman has argued against treating AI systems as conscious entities. In a recent essay, he stated that AI systems do not feel, experience or suffer, and warned against granting technological systems rights or moral protections.
His position contrasts with the openness expressed by some AI industry figures towards further consideration of machine consciousness.
What is the wider debate over AI and human intelligence?
The discussion extends beyond technology companies into broader ethical and philosophical questions about the relationship between human intelligence and machines. During a sermon at St Peter’s Basilica in Vatican City on Thursday, Pope Leo XIV emphasised the distinction between rapidly processing information and understanding experiences through human life and meaning.
Anthropic’s policy update adds to the debate over how AI systems should respond to user behaviour, while leaving the question of machine consciousness unresolved.