Anthropic to start embedding invisible watermarks in Claude’s AI-generated text—as the industry scrambles to police AI slop
New models will embed an invisible, machine-readable signal in AI-generated text, part of an industry-wide push to distinguish human writing from AI output.
Anthropic is set to introduce a subtle watermark that can help identify AI-generated text in its Claude models. The watermark will be incorporated starting August 2, making it machine-readable and imperceptible to humans. This move follows new EU AI Act transparency rules that came into effect on the same date. The watermark will be embedded within the model level and will remain consistent across different platforms like the chatbot, API, and tools like Claude Code.
Images generated by Claude will also have a watermark, alerting users if the file has been tampered with. While text watermarking is more challenging than image watermarking, Anthropic believes this move will help differentiate human-generated content from AI-generated content. This development comes amidst a broader industry effort to address the issue of low-quality AI-generated content and the EU's efforts to regulate AI-generated content.
Written by urgent.news from Fortune's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
