Exploring AI Competencies and Training Needs for Research Infrastructures: Insights from a Knowledge-Exchange Workshop

On Thursday, 27 February 2025, RItrainPlus organised a workshop in the frame of the Staff and Knowledge Exchange programme. This interactive workshop was co-organised by ELIXIR/EMBL-EBI, the Biodata Developers (BioDev) Network and the Wellcome Sanger Institute at the Wellcome Genome Campus in Hinxton (United Kingdom). The focus of the workshop was on ‘AI Competencies: challenges and needs for research infrastructures and core facilities’, bringing together 22 professionals from 11 institutions and 9 countries to explore the competencies necessary to support the use of Artificial Intelligence (AI) in research environments.

The day began with a tour of the Data Centre on campus, where participants learned about the infrastructure that supports high-performance computing and the latest technologies used in research. This set the stage for the workshop, which was designed to be highly interactive, with participants engaging in discussions, group work, and short presentations throughout the day.

The first session introduced participants to AI-related personas—exemplar job titles and responsibilities—helping to ground discussions in real-world roles. Participants then broke into three groups based on their backgrounds—managers, researchers, and engineers—to build on the personas presented. This exercise led to insightful conversations which refined and even created new personas, identifying the key responsibilities and competencies needed to support AI research. The breakout sessions gave everyone the opportunity to actively contribute and shape the direction of the discussions.

After lunch, the group came together to hear two flash talks. The first, from the European Infrastructure for Translational Medicine (EATRIS), discussed the challenges faced by the research infrastructures in training on the use of AI. The second, focused on the BiotrAIn project, introduced a curriculum designed to support AI education for bioscientists in Latin America. Both talks sparked lively discussions, with participants sharing their experiences and ideas, setting the stage for potential collaborations.

Following the flash talks, the participants returned to their breakout groups to focus on the training needs for specific competencies, reviewed existing training at their institutions, and pinpointed the common gaps that need to be addressed. The groups then discussed the challenges of providing the necessary training, such as overcoming resource limitations and adapting to rapidly evolving AI technologies.

The workshop then shifted to a large-group discussion, exploring strategies to address the identified training gaps. Two more flash talks were presented, on how to use ChatGPT to generate bioinformatics training ideas and the challenges of integrating AI tools into reproducible, user-friendly pipelines. This session emphasized the need to build foundational AI literacy within research teams and focus on reproducibility before delving into more advanced AI techniques.

The final session of the day addressed sustainability, a topic of growing importance in AI discussions. Participants debated the dilemma of energy-efficient solutions being labeled as ‘sustainable’, when they can often lead to as much, if not more, energy consumption (albeit with increased productivity), and thus have a comparable impact on the environment. The conversation sparked a critical dialogue about how institutions could meaningfully reduce their environmental footprint while still embracing AI innovations.

The workshop wrapped up with a summary of the day’s discussions, followed by a dinner at the Red Lion pub in Hinxton, where participants continued their conversations in a more informal setting. 

Feedback from attendees was overwhelmingly positive, with most rating the workshop as a ‘very good’ or ‘excellent’ experience. More than 80% of participants said they would recommend the event, and many highlighted the value of the discussions, especially around AI competencies and training needs. The interactive format, which allowed everyone to contribute and share ideas, was also widely appreciated. The opportunity to reflect on how these conversations would impact their work and institutions was seen as a valuable takeaway for future workshops.

In conclusion, the workshop successfully created a space for participants to explore the staff competencies required to integrate AI into research infrastructures, discuss the training gaps that exist, and collaborate on solutions to address these challenges. It provided valuable insights and fostered new connections, setting the stage for continued collaboration in the rapidly evolving field of AI in research.

Some of the feedback from participants included:

“Fantastic session bringing like-minded individuals together to share their enthusiasm for an important societal challenge.”

“Very current themes, highly relevant to different stakeholders.”

“Well thought-out sessions that flowed from each other, excellent event organisation, and great to have all the discussions documented.”