From July 8-10, the workshops designed by DIID for pupils on the topic of “AI & Discrimination” took place for the first time. A total of three school classes of 25 pupils each from Düsseldorf's Görres-Gymnasium high school developed knowledge and skills for the responsible use of AI and how these applications can promote discriminatory structures during the five-hour workshop in the university building. The workshop was funded by the Citizens' University staff unit with around €1000.
AI is increasingly shaping our society and will lead to fundamental changes in all areas of life in the foreseeable future. Many AI applications such as ChatGPT are already common tools for young students. Whether these changes will be equally beneficial for all people also depends on successful ethical reflection. The “AI & Discrimination” workshop was created through the conceptual development of the DIID teaching project “Ethical Issues of Artificial Intelligence”. The original format of a 90-minute school lesson was expanded into a five-hour workshop and adapted to the needs of students in grades eight to eleven. In terms of content, the workshop dealt with the concept of discrimination and combined this with teaching the basic workings of algorithmic decisions. Moral decisions were tested in practice using the “Moral Machine” and discussed in the group. In this way, the students' understanding of the opportunities and risks of the growing use of AI and their ability to deal with these systems critically and responsibly was strengthened, a key skill for the future of young people in an increasingly digitalized society.
The workshops were conducted by DIID employees Till Fahnenstich, who was responsible for the content of the workshop, Anna-Maria Linstaedt and DIID coordinator Dr. Dennis Frieß. Didactically, the workshops were based on both face-to-face dialog and digital-interactive formats. Following the example of Rolf Sistermann's tried-and-tested “Bonbon model”, the students were first encouraged to come up with intuitive ideas and then guided towards scientific points of view for solving problems. To this end, the students were given an introduction to the technological foundations of AI and considered which discriminatory mechanisms could underlie them. Philosophical thought experiments such as the “trolley problem” and the “moral machine” served as concrete examples of ethical challenges. The subsequent discussions in plenary on the question of whether or not the students themselves discriminated through their decisions were lively and created starting points for further questions of ethical decision-making.
The final content section of the workshop dealt with the questions of how and why AI discriminates in the first place. In this context, the participants were introduced to the principles of machine learning using the “Teachable Machine”. The practical introduction demonstrated to the students the concrete consequences of systematic distortions in training data. A manipulated data set, which was compiled by DIID member Dr. Katarina Boland in collaboration with HeiCAD and kindly made available to us for this workshop, served as the basis. In the plenary session, the participants recognized the connection between discriminatory tendencies in society and their representation within data sets used to train AI systems.
Each workshop ended with a short quiz to test the knowledge gained in the workshop and a round of feedback to the workshop leaders. Overall, the workshops provided all participating students with new knowledge and skills in dealing with AI. The workshops thus offered added value in terms of content and interdisciplinarity to content already covered in class.
Following the positive results of testing this workshop concept, the aim now is to evaluate the lessons learned from the first phase and develop new formats for dealing with AI. Interfaces with other disciplines outside of philosophy are also conceivable, such as the much-discussed use of AI in online discussions or social media within communication science.
We would like to thank all classes and teachers for their committed and interested participation! We are delighted with the overall success of all three workshops and are looking forward to their future development.