Be Nice to Your AI

When I was at Singularity University, we joked that you should be nice to your AI and robots – once they become sentient, they might remember all the abuse they suffered from your bullying. It made for some cheap laughs. Anthropic just updated its usage policy prohibiting “sustained and needless abusive or cruel behavior” toward Claude. Anthropic labels this as model welfare – a standard we ought to apply to how we treat each other, too.