Agentic artificial intelligence (AI) shows great potential for automating data processing in research contexts and is developing quickly.

Unlike standard chatbots, which respond to single prompts, agentic AI systems can plan tasks, use tools, retrieve information, organize data, and carry out multi-step workflows with varying degrees of autonomy.

In terms of research data management (RDM), agentic AI systems can perform tasks such as renaming and organizing files, completing metadata, identifying gaps and inconsistencies in datasets, converting formats, and drafting documentation, rather than having these tasks done manually.

This introductory workshop is designed for data stewards, network members, and researchers who wish to learn how agentic AI can support research projects, RDM and data curation. The workshop will explain the practical meaning of “agentic”, where these systems may be useful in research, and the risks they introduce.

The session will include a live demonstration of agentic AI, followed by the opportunity for hands-on experimentation in a sandbox environment. Participants will learn how to formulate tasks, evaluate outputs, document AI-assisted processes, and determine when not to use agentic AI.

By the end of the session, participants will be able to:

• identify suitable use cases for agentic AI in research projects
• distinguish between helpful automation and unreliable delegation
• design simple workflows with agentic AI
• assess security risks in research projects when using agentic AI.

The workshop concludes with a discussion of how agentic AI and other AI tools can be used effectively and securely for research projects at the university and the steps needed to achieve this.

Lecturer

• Dr. Olga Serbaeva (Open Science, University Library, University of Basel)

Course Language: English

Data stewards and other members of the university's RDM Network will be invited to this course by email. Other people who are interested in attending the course are asked to contact the coordinators of the Data Stewardship Program (researchdata@clutterunibas.ch).