European Commission updates guidelines for responsible use of generative AI in research

The European Commission has updated its guidance on generative AI in research, adding new recommendations on agentic AI, peer review, and research funding evaluations

The European Commission has published an updated version of its European Research Area (ERA) Living Guidelines on the responsible use of generative artificial intelligence (AI) in research.

Announced on 16 October 2026, the update is the most comprehensive revision since the guidelines were first introduced in 2024. It aims to help researchers and the wider scientific community use generative AI responsibly as the technology continues to develop rapidly.

The guidelines offer practical recommendations to encourage AI adoption while maintaining the integrity and accountability expected in scientific research.

New guidance on agentic AI

One key addition is guidance on agentic AI, AI systems that can operate with greater autonomy and carry out increasingly complex tasks.

As these systems become more capable, questions surrounding human oversight, reliability and accountability are becoming increasingly important.

The updated guidelines aim to help researchers understand these challenges and ensure that using more autonomous AI tools remains consistent with responsible research practices.

Greater clarity on peer review and funding

The revised guidance also provides more detailed recommendations on using AI in peer review and research proposal evaluation.

These activities play an important role in maintaining research quality and determining which projects receive funding. The Commission has clarified and expanded its previous recommendations in response to concerns and needs raised by the scientific community.

The aim is to support responsible AI use in these processes while recognising the importance of accountability and appropriate oversight.

Research integrity remains central

Despite the additional guidance, the guidelines retain their practical and accessible approach. Their revised structure is designed to make recommendations easier for the research community to understand and apply.

Principles such as transparency, responsibility, and accountability remain at the centre of the framework. These principles are intended to guide decisions about when and how generative AI should be used in research.

An evolving framework

The ERA Living Guidelines are designed to evolve alongside AI developments. Since their initial publication in 2024, they have undergone several revisions, with the previous update released in May 2026.

The work involves collaboration with countries and stakeholders across the European Research Area. An open feedback process also allows researchers and other interested groups to contribute suggestions for future improvements.