Generative AI in research: EU updates ERA guidelines

On 16 October 2026, the European Commission published updated ERA Living Guidelines on the responsible use of generative artificial intelligence (AI) in research. The recommendations are aimed at researchers, research organisations and research funding organisations. New areas of focus include agentic AI and the use of AI in peer review and the evaluation of funding proposals.

Generative AI can support scientific work, for example through literature searches, coding or presenting research findings. At the same time, its use raises questions about reliability, transparency and confidentiality. The revised guidelines for the European Research Area (ERA) provide practical guidance grounded in research integrity.

Agentic AI: responsibility remains with people

AI agents can increasingly plan complex tasks and carry them out with a degree of autonomy. The updated guidelines therefore recommend appropriate checkpoints at which researchers review and validate decisions, particularly those concerning methodology. Researchers remain responsible for scientific outputs, including when using autonomous tools.

AI in peer review and the evaluation of funding proposals

Another key focus is research assessment. The guidelines clarify the conditions under which generative AI can provide support. The following principles are particularly relevant:

• Scientific assessment remains a human responsibility: Expert judgement must not be delegated to AI tools.
• Confidentiality must be protected: Unpublished proposals, manuscripts and review reports must not be disclosed through unsecured AI tools.
• Applicable rules must be followed: Reviewers must comply with the requirements of the relevant funding organisation, institution or journal.

The guidelines also provide recommendations for funding organisations on explaining permitted uses of AI and supporting transparency in the review process.

Practical recommendations for research

The ERA guidelines are non-binding recommendations. They complement existing legal, institutional and funder-specific requirements. They provide a basis for researchers and organisations to critically assess AI use, clarify responsibilities and transparently disclose substantial AI contributions.

First published in 2024, the Living Guidelines continue to evolve through collaboration with countries and stakeholders across the European Research Area. 

Do you use generative AI in your research or when working with funding proposals? The updated guidelines offer guidance on transparent and responsible use.

• Download the updated ERA guidelines (PDF, English)

Dipl.-Wirt.-Ing. Dan Gutu
Scientific Officer Information & Communication Technologies |
Engineering & Natural Sciences
Phone: +49 89 9901888-136
Email: gutu@no-spam-pleasebayfor.org