AIHacker Noonabout 5 hours ago
AI Watermarks Arrive, But Prove Little
5 min read
OpenAI embedded SynthID marks into GPT-Live voice output on July 31, and Anthropic began watermarking Claude text on August 2. The marks show AI processing, not authorship; editing, translating, or transcribing removes them. Detection requires a secret key, limiting independent verification.
Level
Hype check
Tap to vote and see what everyone thinks.
