NLBSE 2025

Title: Breaking the Silence: the Threats of Using LLMs in Software Engineering

Abstract: Large Language Models (LLMs) are transforming the Software Engineering (SE) landscape, influencing tasks such as code generation, program repair, and test case creation. While these models promise enhanced productivity and innovation, they also introduce significant challenges. Issues like data leakage, closed-source models, and the reproducibility of results raise critical concerns for researchers and practitioners alike. In this keynote, we will explore the evolving role of LLMs in SE, shedding light on the hidden complexities and potential risks that come with their use. More importantly, we will present practical guidelines for both researchers and industry professionals to ensure rigorous, transparent, and reproducible LLM-based research. By drawing on real-world examples and best practices, this talk aims to spark a broader discussion on how we, as a community, can responsibly harness the power of LLMs for software engineering.

Biographies:

June Sallou is an Assistant Professor in Sustainable Software Engineering @Wageningen University & Research, in the Information Technology research group. Her research interests include Green AI & Software, and Scientific Computing, and I am exploring how to apply Approximate Computing for sustainable (scientific) SE. Looking at the Sustainability in Software Engineering, she aims to help build more sustainable (scientific) software systems, and more specifically Green AI.

Annibale Panichella is an Associate Professor at Delft University of Technology (TU Delft) in the Netherlands. Within SERG, he leads the Computation Intelligence for Software Engineering Lab (CISELab) and focuses on advancing the fields of software testing, search-based software engineering, software testing for AI, and empirical software engineering. He is the track lead for “Testing for and with AI” at the AI4SE, TU Delft's tenth ICAI lab, funded by JetBrains, a global leader in intelligent software development tools. Additionally, he is the principal investigator for the UBRI program at TU Delft, a pioneering joint initiative with Ripple, a leader in the Blockchain sector. He has received two Most Influential Paper Awards at SANER (2024) and ICST (2025) and several Best Paper Awards at ICPC, SBFT, and SSBSE. He serves and has served as a program committee member of various international conferences (e.g., ICSE, ESEC/FSE, ISSTA, GECCO, ICST) and as a reviewer for various international journals (e.g., TSE, TOSEM, TEVC, EMSE, STVR) in the fields of software engineering and evolutionary computation.