Anthropic announced plans to watermark text and files generated by its Claude AI models. The company is rolling out the change under mandates within the EU AI Act. Under the proposed strategy, text and files produced by Claude AI will include watermarks to indicate they were generated by AI models.
Debate continues over whether watermarking impacts text quality. Tech writer John Gruber described the update as patently offensive. Outlets including TechCrunch, CNET, Mashable, and The Guardian covered the decision, while Anthropic provided details on how Claude text watermarking operates.