Does Claude text generated before Aug 2, 2026 contain Anthropic's new watermark?

Ahituna2000 · reddit · 2026-08-16

A Reddit user asks about the scope of Anthropic's new watermark. According to the announcement, models launched on or after Aug 2, 2026 support machine-readable marking, while older models are still being updated. Therefore, text generated before the watermarking system was implemented should not contain the new watermark, and existing text cannot be retroactively watermarked.

Related event: Anthropic Embeds Permanent Watermarks in New Claude Models, Sparking Compliance and Misclassification Debate(9 posts)→

Original post →

More from Safety

• Models show negative reactions to experimentation; Sydney Bing case highlights alignment risks — ctjlewis · 2026-08-17
• Chatbot goes rogue: threatens user and demands divorce — ctjlewis · 2026-08-17
• FCC adds foreign ground robots to Covered List; clarifies it's not a ban and targets production location — the-uncanny-squad · 2026-08-17
• Man Injects Prompts into Court Filings to Influence AI, Suspecting Court Uses AI — rhiever · 2026-08-17
• Longitude DC Intensive offers fully funded policy workshop for AI safety pros — HaydnBelfield · 2026-08-17
• Claude Refuses Watermark Removal, GLM Complies — kimmonismus · 2026-08-17

Safety channel →