Anthropic plans to add an invisible mark to AI text -- as the industry scrambles to police AI slop | Fortune

Anthropic says it plans to add an invisible, machine-readable watermark to AI text it generates, starting with models released on or after August 2. The signal will be woven into Claude output so users can’t see it or notice any effect on quality or readability. It is intended to travel with copied and pasted text, and may survive some editing, though heavy rewriting or translation could remove it. Because the watermark is placed at the model level, it should appear across Claude’s chatbot, API access, and tools such as Claude Code, and Anthropic also plans watermarks for images. The move aligns with the EU AI Act’s transparency requirements that begin August 2. Fortune also notes a broader push against low-quality AI “slop,” including efforts by Substack and YouTube to detect or limit synthetic content.