How Claude's Text Watermarking Works
Future Claude models will embed a watermark in generated text to help determine if the text was AI-generated, complying with the EU AI Act. The method, based on SynthID-Text, does not affect output quality or readability.
How it works
LLMs pick words one at a time, choosing among candidates. For low-stakes choices (e.g., “overcast” vs “grey”), the selection is random. Watermarking uses a key plus preceding words to seed that randomness, creating a detectable pattern. Detectors with the key can check if the word sequence matches the key-based pattern.
Impact on output
Anthropic reports no quality change, and Google DeepMind's SynthID-Text paper found no statistically significant differences in user ratings. Human raters saw no quality difference. No hidden characters, no added tokens, no extra cost.
Compliance
Implemented to meet EU regulations requiring AI providers to mark AI-generated content. Other major providers signed the same Code of Practice.
📖 Read the full source: HN AI Agents
👀 See Also
Why Write Code in 2026: Human Coding Still Matters for AI Agents
Doug Turnbull argues that even with powerful AI coding agents, humans must write code to understand architecture, reduce fragility, and guide agents effectively.
GM Lays Off 600 IT Workers, Hires AI-Focused Engineers for Agent and Model Development
General Motors cut 600 IT employees (~10% of the department) to hire workers with AI-native skills: agent development, data engineering, cloud engineering, prompt engineering.
Claude Code Performance Regression Diagnosed: Configuration, Not Model Intelligence
Anthropic's postmortem reveals Claude Code's performance drop was caused by three product changes — default reasoning effort, session caching bug, and prompt-verbosity — not model degradation. Rollback restored performance.
OpenClaw Mistral Provider Broken Since 2026.3.8 Update, Community Seeks Alternatives
OpenClaw users report persistent HTTP 422 errors with Mistral models since the 2026.3.8 update, with no fixes in subsequent releases through 2026.3.13. The issue affects all Mistral-related functionality while direct API calls work normally.