Anthropic announced a new text watermarking system for Claude on Tuesday, August 11, 2026, aimed at helping meet European Union transparency regulations for artificial intelligence. The system embeds statistical signatures into text output during token generation, modifying how tokens are selected to create a detectable signature for traceability. The initiative is designed for EU transparency compliance and targets text generations from models like Claude 3.5 Sonnet. Anthropic’s CEO Dario Amodei has emphasized safety standards in AI deployment, fitting this governance theme. European Union officials outlined compliance deadlines for generative AI providers, creating regulatory pressure that pushes safety and transparency mechanisms into production roadmaps. The watermarking allows regulators and auditors to detect the signal after text is produced, potentially reducing moderation ambiguity and documentation workload for builders, enterprises, and platform teams. OpenAI previously discussed cryptographic provenance initiatives for ChatGPT outputs in late 2023.