Anthropic’s move to watermark Claude outputs is already producing backlash from some users who wanted their AI use to stay hard to detect.
TechCrunch reports that Anthropic is inserting invisible markers into Claude-generated editorial text to help identify it as AI-generated. The company rolled out the policy in connection with the EU AI Act’s Transparency Code, which pushes technology companies to label AI-generated or AI-edited content in ways computer systems can detect.
Some users have complained online that watermarking could expose chatbot use in classrooms or workplaces. That reaction shows the tension around transparency: the same signal that helps institutions identify AI-generated text also makes it harder for people to pass off machine-written work as unaided human writing.
The practical limits are still important. Watermarks may be weakened by paraphrasing or heavy editing, and detection systems can produce disputes. Even so, the change points toward a future where major AI providers build provenance features into everyday text generation rather than leaving disclosure entirely to users.