OpenAI to Watermark ChatGPT Outputs by Default in the EU, Despite Reliability Gaps
OpenAI is preparing to watermark ChatGPT-generated text by default for users in the European Union, a move that reflects intensifying regulatory pressure over AI transparency. But the measure is unlikely to be a robust safeguard: watermarking systems remain imperfect, and the markings can often be stripped or bypassed with modest effort.
