Anthropic Rolls Out Checker for Claude Content

Brian Roemmele posted about Anthropic releasing a content checker and applying C2PA metadata to files that Claude generates or touches. He also noted invisible statistical watermarks woven into the model's text. Replies from Colin Glassey and others stated that Anthropic has no moral right to claim ownership of LLM outputs. They argued a machine cannot hold copyright and that the text draws from sources users could access directly.