Abstract

Artificial Intelligence (AI) is rapidly transforming the landscape of software engineering (SE) with bold promises of unprecedented productivity and innovation. While industry giants promote AI-driven tools as game-changers, recent studies paint a more nuanced picture, with some contradicting the hype. This keynote critically examines the current state of AI in SE from a practitioner's perspective. We will explore the optimistic narratives promoted by major tech companies, confronting them with the realities revealed by recent empirical research and identifying realistic opportunities for integrating AI into everyday software development. Most importantly, the keynote will address the pressing questions that practitioners have for researchers. What problems truly need solving? By bridging the gap between research and practice, this talk aims to spark a dialogue that will contribute to shaping the research agenda for AI in software engineering.

Bio

Andreas Jedlitschka leads the Data Science department at the Fraunhofer Institute for Experimental Software Engineering. His department deals with a systematic approach to data-related issues (why, what, when, where, who, which, how). His research interest is in empirical and data-driven decision support, where he has published more than 70 papers in journals and conferences. More recently, Andreas has focused his work on using (generative) Artificial Intelligence (AI) within the software engineering life cycle. From a research perspective, he is particularly interested in exploring AI use cases for the benefit of humanity and in the empirical evaluation of AI in setting with Human-AI-Collaboration. Another important aspect is the trustworthiness of the results which can be obtained from AI (in particular Generative AI). Andreas is co-author of a German patent on "Method for unobtrusive monitoring of a person and system for carrying out the method". His most cited paper is on "Guidelines for Reporting Controlled Experiments in Software Engineering". The guidelines provide a starting point for the adoption of the evidence-based SE paradigm. Dr Jedlitschka is a member of several programme committees (e.g. ESEM, PROFES) and served as Programme Co-Chair of ESEM 2016 and as General Chair of PROFES2013 and General CoChair of PROFES2023. In addition, he had several other roles in conferences, e.g., he organized CESI and WSESE Workshops at several ICSE conferences and significantly supported the organizers of Dagstuhl-Seminar 06262. He represents Fraunhofer IESE in the International Software Engineering Research Network (ISERN) and is a life-time member of the ISERN Steering Committee. In addition, he is elected five-year member of the ESEM Steering Committee. He has had leading roles in projects with industry and in the European FP5 (ESERNET), FP6 (EMERGE, VIDE, ReDSeeDS), FP7 (OASIS, AALIANCE, MUNDUS) and H2020 (Q-Rapids), ITEA (GreenCode). Andreas was a lecturer for "Empirical Modelling and Methods" and “Software Project and Process Management) at the Rhineland Palatinate Technical University of Kaiserslautern and gives regular invited talks on Artificial Intelligence.Andreas Jedlitschka received his M.S. (Diplom, 1994) and Ph.D. (Dr.-Ing., 2009) in Computer Science from the University of Kaiserslautern. After seven years as an IT consultant at Voss&Kamb und Partner GmbH, he joined Fraunhofer IESE in 2000 as a project manager and scientist.