EU's push to watermark AI writing is a bad idea that will hurt output quality, argues dev

VraserX · x · 2026-10-06

VraserX criticizes the EU's requirement that OpenAI and other AI labs watermark AI-generated writing, calling it an awful idea.

The core argument: you cannot deliberately manipulate a model's writing patterns to make them detectable without paying a cost in output quality. The take highlights the fundamental tension between detectability and generation quality in watermarking policy.

Related event: OpenAI launches textGrain text watermarking, starting in the EU(25 posts)→

Original post →

More from Safety

Safety channel →