OpenAI to watermark ChatGPT text in EU under AI Act compliance

OpenAI, along with Google, Anthropic, Meta, Microsoft and Mistral, has committed to the European Union's AI-content code of conduct. As part of this commitment, OpenAI will introduce watermarking on ChatGPT-generated text across the EU.

The watermark aims to help identify AI-generated content under the EU's AI Act framework. However, testing has revealed that the watermark can deteriorate significantly after text editing, potentially limiting its effectiveness in detecting manipulated outputs.

The move reflects OpenAI's alignment with EU regulatory requirements on AI transparency and content identification.