OpenAI is taking a new transparency step in the EU by adding watermarks to text generated by ChatGPT and Codex. The change is designed to align with the European Union’s AI Act, which emphasizes clearer disclosure around AI-generated content.
For users, institutions, and online platforms, watermarking could make it easier to understand when text may have been produced by an AI system. That matters in settings such as education, publishing, software development, and digital communication, where trust and provenance are increasingly important.
Why this is a win
- More transparency: Watermarks can help people identify AI-generated outputs more reliably.
- Real-world deployment: This is not just a research concept; it is being applied to widely used AI tools.
- Regulatory alignment: The move shows major AI providers adapting to public-interest rules around disclosure.
OpenAI has also acknowledged an important limitation: editing AI-generated text can make invisible marks harder to detect. Even so, the rollout represents progress toward a healthier AI ecosystem where powerful tools are paired with stronger accountability and clearer signals for users.