Anthropic rolls out new ai text watermarking system on Claude – here’s what to know

Anthropic has rolled out an invisible text watermarking system for its Claude AI model, designed to subtly influence word choices using a cryptographic key to create detectable patterns, in compliance with new European AI regulations. The company plans to deploy the technology globally across all Claude responses, despite widespread user criticism that it compromises output quality. Anthropic clarified that light editing or proofreading will not trigger false positive detections.

Anthropic rolls out new ai text watermarking system on Claude – here’s what to know Anthropic /topic/anthropic has detailed how its new invisible watermarking system for Claude /topic/claude text outputs operates following widespread user criticism.- The system subtly influences low-stakes word choices using a cryptographic key to create a pattern detectable only by specialised software. - Introduced to comply with new European AI /topic/ai regulations, Anthropic plans to roll out the watermarking technology globally across all Claude responses. - The company clarified that light editing or proofreading will not generate enough altered words to trigger a false positive watermark detection. Tech /topic/tech commentators and users remain critical of the update, arguing that altering model outputs for compliance compromises the quality for end users.