Anthropic Details Invisible Claude Watermarking After User Backlash

Anthropic has explained in more detail how its new Claude watermarking system will work after users criticized the move and some reported canceling subscriptions. The company says text generated by newer Claude models will carry an invisible, machine-detectable signal that travels with copied or lightly edited text, while not changing the meaning, quality, or readability of the output. The backlash has focused on concerns that human-written or heavily revised work could still be flagged as AI-generated, raising worries for students, writers, lawyers, and other users. The change also matters because it reflects growing pressure on AI companies to make machine-made content easier to identify while balancing transparency with user trust.