In August last year, the company announced that it would allow Claude to end conversations with “harmful” users as part of its “model welfare” research. Anthropic’s policy now allows AI models to refuse to communicate with users who repeatedly treat them cruelly without any apparent purpose.
The current update also prohibits the “improper” use of Claude—for interfering in elections, manufacturing weapons, and surveillance.
“Claude may not be used to make decisions or recommendations about whom to investigate, arrest, or charge within law enforcement or the criminal justice system. We also prohibit using Claude to create or improve tools intended for surveillance,” Anthropic said.
The new rules will take effect on November 12.
Our channel: Node of Time EN