The AI Disclosure Handbook: Building transparency in AI-assisted research

Generative AI is opening new possibilities across the research cycle, from literature reviews and coding to data analysis and writing. As its use grows, so does the opportunity to show clearly and confidently how AI contributes to scientific work.

The AI Disclosure Handbook project is working to make this easier. Its foundation is a co-design study with 26 experts from research, funding and publishing institutions, including the European Commission and its executive agencies. Through focus groups and interviews, the participants shared their needs and co-created early designs for AI disclosure. Building on these insights, the project is developing a flexible approach that each field and institution can adapt to its own practices.

What the project is developing

• The AI Disclosure Handbook is an emerging seven-step methodology for designing AI disclosure. It guides institutions in setting goals, mapping research tasks, defining levels of AI involvement and choosing the right format, so that disclosure is meaningful, consistent and easy to understand. The Handbook will be released by the end of 2026.
• The AI Disclosure Tool (beta) is a free web tool, now open for testing and continuing to evolve. It lets researchers create a disclosure statement in minutes: a simple declaration, a visual overview, or a detailed disclosure. It shows whether AI supported the intellectual contribution or played an assistive role. It integrates the AI policies of major publishers and exports to text, LaTeX, Word and XML.

At its heart, the project treats AI disclosure as an opportunity for methodological transparency, shared learning and reflection on how AI is shaping research. As it moves forward, it aims to grow alongside the research community and help build a culture where responsible AI use can be stated openly and with confidence.

Explore the beta tool