Release of Science Mapping Tool as part of the JRC Scientific AI Hub

The Commission released its AI in Science Strategy last year, setting out plans to enable scientists across Europe to take advantage of the monumental potential offered by AI technologies. 

After contributing to the development of the strategy, the JRC now supports its implementation, partly by hosting a space to showcase activities, studies and indicators on AI in Science: The Scientific AI Hub. The Hub offers visitors the chance to explore JRC’s research with and on AI, and it will gather input from across the research community to become a useful common space. 

Tools to support responsible deployment of AI  

To encourage scientists to fully explore the potential of AI for their research, it is important that they can access support to interpret potential legal requirements or restrictions. That is particularly salient now, as the enforcement of the EU AI Act has begun. When it comes to legal compliance, the EU AI Office hosts an Information Platform, which any stakeholder can use to help determine if they are subject to legal obligations and understand steps needed to comply.   

For the JRC, a key focus is on the needs of the scientific community to confidently and responsibly use AI for research purposes. To provide a helping hand, we have launched an AI Science Mapping Tool, which provides an overview of research relevant to the AI Act.

The tool gives users three primary topics for exploration: research related to AI practices prohibited by the AI Act, research related to AI systems classified as high-risk by the regulation, and research related to the transparency obligations that apply to AI systems deployed for direct interaction. Once a topic is selected, users find a word cloud of the most common research topics within this category, a Sankey showing in which countries topical research has been published, as well as the top 5 most cited publications, the most cited publications, and the venues and institutions with most publications on the topic. 

In this way, the tool maps research communities around the risk levels in the AI Act, and gives researchers an opportunity to explore how their own research may be linked to these levels. 

Community engagement and further development

In October, both the Science Mapping Tool and the Information Platform was presented at the AI in Science Summit 2026 in Dublin. Members of JRC and AI Office teams were there to engage directly with the scientific community, to present the tools as they are, and to understand feedback and needs for future development. 

This also led to a discussion of participants’ perception of the current impact of AI in their respective scientific fields. There was wide agreement that AI is indeed transforming fundamental science, from the understanding of protein structures to weather forecasting. However, it was also highlighted that bottlenecks still exist, they have merely shifted from the stage of generating answers to the stage of verifying them. Participants strongly emphasised that rigour in this process is crucial, as science can only move at the speed at which it maintains trust. 

As for the next steps to foster productive and responsible use of AI for science in Europe, the group hoped that we will build on the European strengths of open data, shared benchmarks, and open models. Success will depend on building major European computing infrastructure, and achieving a combination of technological sovereignty and international openness.  

Several other topics were also discussed by speakers from the Commission, including the Resource for AI in Science (RAISE) initiative, which is intended exactly to develop and build on European strengths. It will work as a virtual research institute for European research on and with AI, pooling computational power, data, talent and research funding.