⚖️Anthropic Bans Sustained Abuse of Claude, Election Fakery
TL;DR
Anthropic updated its usage policy to bar persistent cruelty toward Claude and deceptive election campaigns. Ordinary frustration, dark fiction and model testing remain allowed.
Anthropic updated its usage policy to bar persistent cruelty toward Claude and deceptive election campaigns. Ordinary frustration, dark fiction and model testing remain allowed.
Key Points
New "Do Not Undermine Democratic Processes" section covers voter deception and election disruption
Fake accounts and fabricated news outlets run through Claude are prohibited
The abuse rule targets users who "repeatedly act cruelly toward our models, with no discernible purpose"
Claude was already trained to end persistently abusive conversations; the policy now forbids pursuing them
Why It Matters
This is a usage policy that protects the model, not only people. It is also a signal on model-welfare thinking from a lab that is hiring a lead for 2028 election engagement.
Quick Facts
Comments
Be the first to comment
Enjoyed this article?
Get it daily. 7am. Free. Reads in 5 minutes.
Join 3,558 builders reading daily.