Anthropic Ships Invisible AI Watermarking to Catch Cheating Students

Anthropic has rolled out an invisible watermarking system for Claude AI that weaves imperceptible signals throughout any generated text, making it easy to detect AI-generated school essays. The watermark travels with copied text and persists through some editing, applying at the model level across all Claude products. This move is part of compliance with new European transparency laws, though it will be applied worldwide. Anthropic joins OpenAI and Google in developing text...