Anthropic Is Quietly Watermarking Every Claude AI Output. Builders Are Already Trying to Break It
Anthropic has begun embedding an invisible, machine-readable watermark in all outputs from its latest Claude models, without revealing the method—prompting developers to attempt cracking it.
Anthropic is secretly embedding an invisible, machine-readable watermark into every text output from its latest Claude AI models. The watermark is designed to trace AI-generated content, but the company has not disclosed how it works. Developers and builders are already attempting to reverse-engineer and break the watermarking system. Anthropic has not publicly commented on the initiative, which appears aimed at combating misuse and verifying AI authorship.