Anthropic's New Policy Bans 'Cruel Behavior' Toward Claude

Is this basic decency or is it confused thinking that treats software as a moral patient?

Published 5 hours ago

Anthropic's New Policy Bans 'Cruel Behavior' Toward Claude

story

Above: A smartphone displaying the Anthropic logo and its Claude AI assistant. Image credit: Imen Ben Youssef/Hans Lucas/AFP/Getty Images

The Spin

Pro-industry narrative

Anthropic's latest rule is about basic decency. Chatbots shouldn't facilitate the worst, most callous impulses, which is why the company has imposed a narrow ban on relentless, purposeless cruelty. This costs honest users nothing and doesn't restrict users' frustration, dark creative writing or research.

Anthropic

Industry-critical narrative

Calling a statistical model a possible moral patient that needs shielding from distress is a confused way of thinking. How can Anthropic warn that AI could doom humanity one moment, then fret over whether chatbots suffer the next? Software doesn't have feelings, and policing how people type into a chat window swaps clear-eyed risk assessment for sentiment.

Michael Shellenberger on X