Physix Frontier · News Briefing Card (TheVerge · Oct 9, 2026)

Anthropic Updates Usage Policy, Bans Sustained Abuse of Claude

KEY FACTS

Anthropic updated its usage policy for the first time in over a year, adding high-risk abuse clauses covering election interference, weapons development, and surveillance.
The new policy prohibits users from subjecting Claude to sustained and pointless abuse or cruelty, and allows the company to terminate conversations in extreme cases.
Anthropic consolidated scattered existing restrictions into new rules banning deceptive commercial or political propaganda.
The weapons development ban now extends to weapons software and components, as well as autonomous vehicles such as armed drones.
The surveillance ban explicitly prohibits tracking people without their consent and using Claude to decide who law enforcement investigates.

PHYSIX OBSERVATION

By writing "be nicer to the model" into its usage policy, Anthropic has taken another step ahead on model welfare. But who defines an "extreme case," and whether there are account bans beyond ending a conversation, remain vague. This looks more like a values statement with limited practical enforcement, yet it could push the industry to re-examine the boundaries of human-machine interaction.

Source: TheVerge report