Anthropic Adds Invisible Watermark to Claude to Meet EU AI Act Requirements

Key Points

• Anthropic will embed an invisible watermark in all Claude model outputs, including Claude Code.
• The watermark complies with the EU AI Act’s transparency requirements.
• Anthropic plans to launch a text detection API for users to identify watermarked content.
• EU guidelines state that users should know when they encounter AI‑generated text to prevent misinformation.
• Non‑compliance with the AI Act can result in fines up to €15 million or 3 percent of global revenue.
• Regulators expect AI firms to develop reliable, interoperable watermarking methods.
• The watermark does not affect the meaning, quality, or readability of Claude’s responses.