On Tuesday, January 19, 2021, we discussed the results of our FairandGoodADM project with politicians, experts and practitioners as well as a great number of interested viewers.
Over the last few years, Georg Wenzelburger and Kathrin Hartmann, together with scientists from the fields of computer science and philosophy at the TU Kaiserslautern, drafted guidelines for the use of algorithmic decision making (ADM) systems by state actors. The main focus was on how to ensure the fairness and quality of said algorithms.
An example was given by Univ.-Prof. Dr. Michael Wagner-Pinter from Synthesis Forschung GmbH in a keynote speech: For the Austrian labor market services he developed an ADM system designed to classify jobless persons according to their labor market chances and their resulting support needs.
In two panels, the results of the project as well as the Austrian labor market service example were discussed. The first panel, consisting of Georg Wenzelburger, Carla Hustedt (Bertelsmann-Stiftung), Dr. Pascal König (TUK) and Dr. Hannah Ruschemeier (Center for Advanced Internet Studies, CAIS), exchanged views on ethical and legal aspects of the proposed guidelines.
The second panel, which included Prof. Dr. Katharina Zweig (TUK), Marco-Alexander Breit (BMWi), Saskia Esken (MdB, SPD party leader), Dr. Stefan Heumann (Stiftung Neue Verantwortung) und Tabea Rößner (MdB, Grüne), discussed the need for political regulation and possible impacts on innovation climate.
A consensus was reached by the panelists that an interdisciplinary evaluation is necessary before the implementation of an ADM system – the guidelines that were formulated in the course of the FairandGoodADM project are a good starting point.