Anthropic has told California's governor that courts should be able to block AI models considered risky. No further details of the message were given in the report.