Institute for Information Systems at DFKI wins 1st place in the Science Dialogue

Lea Mayer, Nijat Mehdiyev, Oliver Gutermuth, and Prof. Dr. Peter Fettke won the first place with their contribution "Responsible AI – The use of artificial intelligence (AI) in public administration through explainable and fair AI methods." Their concept for the responsible, transparent, and socially fair use of artificial intelligence technologies in public administration convinced the jury and the audience the most.

Oliver Gutermuth, M.Sc., Lea Mayer, M.Sc., Head of Research Department Prof. Dr. Peter Loos, Deputy Head Prof. Dr. Peter Fettke, and Nijat Mehdiyev, M.Sc., are pleased about the award.

The winning projects in the fields of AI, age mobility, and cybersecurity were already presented on June 16 at the event "Zukunftskongress-Digital." The patron of the science competition, Minister of State for Digital Affairs Dorothee Bär, announced the winners via live streaming. The Science Dialogue is a cooperation of Wegweiser Media & Conference GmbH and ]init[ AG in partnership with the National E-Government Competence Center. It aims to provide a stage for innovative research contributions and to pave the way to practical application.

The project idea of the four scientists of the Institute for Information Systems (IWi) at DFKI aims at designing concepts for the responsible, transparent, and (socially) fair use of artificial intelligence technologies in public administration. To this end, an approach is to be developed that explains the results of an AI-based data-driven decision tool for internal and external stakeholders in public administration while ensuring the fairness of the algorithms used. The goal is to strengthen acceptance and trust in AI.

With innovative technologies, authorities can work in a more modern and efficient way. AI-based consulting systems already support a wide range of processes within public administration, including tax calculation and fraud detection. To exploit the enormous potential, especially in the area of taxation, the complex organizational structures and heterogeneous system landscapes in public administration must be taken into account. The solutions used should be understandable and transparent for employees and the public.

The core of the project is the research, conception, and development of a responsible artificial intelligence (RAI) to support data-driven decision making in public administration. RAI represents an AI-based approach that uses transparent technical mechanisms with explainable results and guarantees the fairness and objectivity of AI algorithms. AI should make fair decisions that do not discriminate against persons based on race, color, religion, or gender.

The business information scientists at DFKI want to create an AI explanatory system that incorporates findings from Explainable Artificial Intelligence (XAI), cognitive sciences, information systems, and human-computer interaction and enables interactive communication between humans and AI applications. To this end, they develop methods and tools that reflect not only the transparency of the technology but also the technical reference, in order to ensure the fairness of the underlying algorithms and to convey this to the outside world.

Further Information:
https://www.zukunftskongress.info/de/science-dialog

Video of project pitch (German):
https://www.youtube.com/watch?v=wKTyVz79dUg

Slides Responsible AI (German):
https://cloud.dfki.de/owncloud/index.php/s/6r3WapBLmNfwmMJ#pdfviewer

Share this post:

Contact:

Prof. Dr. Peter Fettke

Deputy Head of Research Department, DFKI

German Research Center for Artificial Intelligence
Deutsches Forschungszentrum für Künstliche Intelligenz