OpenAI Supports EU Code of Practice on Transparency of AI-Generated Content

The European Union's push for transparency in AI-generated content has gained momentum with OpenAI announcing its support for the recently published Code of Practice. This move builds upon years of internal research and development by OpenAI to strengthen provenance for AI-created content.

As people increasingly rely on AI tools to create, edit, and share online content, it's essential that users have context about what they see. The EU Code of Practice aims to address this issue by promoting transparency in the digital ecosystem.

OpenAI has been working towards this goal since 2024 when it began adding C2PA metadata to its DALL-E 3 image generation tool. Since then, the company has continued to improve provenance through enhanced marking and detection methods, new research, and the release of a public verification tool.

The EU Code of Practice is an important step in implementing the EU AI Act, which seeks to build a more transparent digital ecosystem. OpenAI's support for this initiative reflects its commitment to responsible AI development and governance in Europe.

OpenAI has been actively involved in developing interoperable standards with other stakeholders through coalitions like C2PA. This collaboration aims to advance provenance techniques that can be used across the entire value chain, from content creation to sharing online.

The company's support for the EU Code of Practice is part of its broader approach to AI governance in Europe. In 2025, OpenAI became the first US company to sign the EU's General-Purpose AI Code of Practice, demonstrating its commitment to clear and workable rules that promote responsible AI development.

Provenance signals can help users understand where content comes from, how it was created or edited, and whether it is authentic. This information can also aid in detecting disinformation campaigns and supporting election integrity.

While provenance methods are useful, they still have limitations. Metadata can be removed, watermarks can degrade, and labels only provide context when people encounter the content directly.

The EU Code of Practice sets a high level of ambition for advancing provenance techniques. As it is implemented, it's essential that these methods remain grounded in practical solutions and flexible enough to reflect current technical limitations.

OpenAI will continue to strengthen content transparency across its products, contribute to interoperable standards development, and improve verification tools. The company looks forward to collaborating with the AI Office, Member States, and the wider AI community to build a safer, more transparent digital ecosystem.