LLMs are currently used in many areas of database technology. Their applications are highly diverse and include design tasks, programming, query generation, optimization, data integration, and more. The results are often remarkably good and surprisingly accurate. However, all application areas share one important limitation: we still cannot rely on LLM outputs to be fully trustworthy, reproducible, or stable over time.
This means that LLMs are very well suited for building prototypes, but they are not yet ready for developing robust, industry-grade systems. This challenge will be the central focus of the Spring Meeting of the GI Special Interest Group on Database Systems (Frühjahrstreffen der GI-Fachgruppe Datenbanksysteme). The following three aspects will be the main focus of the Spring Symposium.
We want to examine three central aspects:
LLMs in the database domain:
- Application scenarios for LLMs in DBMS and data engineering processes
- Current capabilities and limitations of LLMs
How can we make LLM results reliable?
- Methods for validating, verifying, or cross-checking LLM outputs
AI in Education
- Development of teaching formats and environments that teach and assess essential skills, such as academic writing
- Identification of new skills that computer science students need; teaching of these new skills in data engineering lectures
Format of the Symposium
Invited presentations will be given at the Spring Symposium, and further presentations will be selected from submissions. There will also be a poster session. We invite researchers working in this field to submit a presentation proposal or poster. Presentations can be submitted until January 9, 2026 (AoE). Posters can be submitted until January 23, 2026 (AoE).
Further information (about the symposium location, recommended hotels, the submission process, and the program) can be found here: www.go.ur.de/fgdb-spring-symposium-2026 (external link, opens in a new window)