OpenAI will watermark ChatGPT outputs by default—but only in the EU

OpenAI has announced that it will begin automatically watermarking text generated by ChatGPT for users within the European Union. This initiative is designed to ensure compliance with the EU AI Act, which mandates the clear identification of AI-generated content. While the feature will be available globally, it will remain disabled by default outside of the EU. The company's proprietary watermarking method, dubbed textGrain, embeds subtle patterns into word choices that are imperceptible to humans but detectable by specialized tools. OpenAI plans to provide access to these detection tools to a select group of researchers and organizations, with a broader approval process to follow. Despite these efforts, experts note that such watermarking techniques remain relatively easy to circumvent, raising questions about the long-term effectiveness of the technology in preventing AI-generated misinformation.