Our approach to EU text provenance rules

OpenAI has published a blog post detailing their approach to implementing the European Union's text provenance rules, which require companies to provide information about the source of text generated by AI models. The rules aim to increase transparency and accountability in AI-generated content. OpenAI's approach involves adding a metadata field to their models' output, which will include information about the model, its training data, and the specific prompt used to generate the text. This information will be available in the model's output and can be accessed by users. The company is also providing a tool for developers to verify the provenance of text generated by their models.