Anthropic adds EU watermarking, Stratechery calls it a 'terrible idea'
Anthropic's EU-mandated watermarking may be worse than it seems, argues Stratechery.
Deep Dive
Anthropic is adding watermarking in response to the E.U.’s AI law — and Stratechery calls it a terrible idea, first and foremost for philosophical reasons.
Key Points
- Anthropic is adding watermarking to comply with the EU's AI law, per Stratechery.
- Ben Thompson calls the move 'a terrible idea' for philosophical and technical reasons.
- The technique likely embeds invisible statistical markers in AI text, risking quality and trust.
Why It Matters
EU AI law is forcing real design trade-offs on frontier models; enterprises must prepare for watermarked outputs.