The AI developer’s latest usage policy prohibits “sustained and needless abusive or cruel behavior” toward its models.

Anthropic HQ:

“computer, say ow”

ow

“oh my god what have i done”