AI for All Three universities jointly introduce AI into teaching
Artificial intelligence is considered one of the central key technologies of the future. For this reason, the topic of artificial intelligence is to be increasingly established in teaching and further education at universities and colleges. Technische Universität Braunschweig convinced a competition jury with a joint project in which TU Clausthal and Ostfalia are also involved. The project is being funded for four years as part of the Bund-Länder funding initiative “Künstliche Intelligenz in der Hochschulbildung” (Artificial Intelligence in Higher Education).
Everyone is talking about AI – in many areas of life, users are hoping for simplifications of complex processes. “Nowadays, AI plays a role in more and more areas of life and is a permanent fixture. AI allows an improved user experience in many applications of speech or image recognition, such as in voice assistance systems or automated driving, but also in the personalisation of social media or in educational software,” says Professor Tim Fingscheidt, who heads the Department of Signal Processing and Machine Learning at TU Braunschweig.
Reliability and Trust in Results
When AI becomes such an important part of everyday life, users should be aware of how to use the new technology. “To know how and where to use software responsibly, you should understand the method behind it,” says Professor Sebastian Stiller from the Institute for Mathematical Optimization. He says there are significant differences between the methods that are now referred to as AI. “Some methods work with clearly defined and exact criteria – like a pocket calculator. Others deliver results that are only valid in a statistical sense. Imagine urban planners developing concepts based on AI tools whose validity they cannot judge. We want our graduates to understand what they are using and therefore know what they are claiming.”
Professor Tim Kacprowski, Head of Data Science in Biomedicine at the Peter L. Reichertz Institute for Medical Informatics, is also convinced that methodological knowledge in teaching helps both to facilitate the interpretation of results and to increase confidence in AI. “We need to talk much more intensively about the possibilities and limits of AI in teaching and give all potential users the opportunity to acquire basic methodological knowledge.
How AI should be introduced into teaching
This is where the joint project “KI4ALL” comes in. “First of all, we will carry out an status analysis and create a cross-location AI lecture directory. The term AI is very broad. It includes various methods that are used in different subject contexts. This subject diversity has also motivated our research proposal: On the one hand, we want to create basic offers together in the sense of an AI literacy in order to make all target groups ready for the topic. On the other hand, the individual locations have different subject foci in research, whose future viability we want to strengthen by specifically integrating subject-specific AI methods into teaching and further education,” says Professor Henning Wessels, Junior Professor for Data-Driven Modelling and Simulation of Mechanical Systems, about the approach in the research project.
AI course catalogue and digital teaching offers
The cross-locational can be imagined as a new database or website with an interactive interface where AI teaching offers can be found and booked for all participating locations – TU Braunschweig, TU Clausthal and Ostfalia.
The basic content mentioned includes teaching-learning content to achieve a kind of AI competence or also “AI literacy” for a broad audience – students, employees in administration, further education and industry. Different formats are planned, including online lectures and computer labs in which independent programming tasks must be solved within the framework of neural networks in order to also gain practical experience with AI procedures.
Only memorised solutions?
How reliable and robust is an AI anyway? Do the algorithms really fit a concrete problem or has the software merely “learned by heart” the solution to the problem from another source? Such questions can be explored, for example, in intra-year competitions. Another focus will be on so-called federated machine learning, in which several smaller instances of artificial intelligence are combined into one large AI: This will make it possible to provide a very large, cross-location data base without the final machine learning model ever having access to the primary data. New opportunities arise here in terms of both the use of data and computing resources as well as data protection.
About the Funding Initiative
40 individual projects of higher education institutions and 14 collaborative projects of several higher education institutions are being funded within the framework of the Bund-Länder funding initiative “Künstliche Intelligenz in der Hochschulbildung” (Artificial Intelligence in Higher Education). The funding initiative, which has a budget of around 133 million euros and reaches 81 higher education institutions across Germany, aims to establish AI broadly in the higher education system. On the one hand, funding is provided for measures that contribute to the qualification of future academic specialists. For example, universities are supported in the development of study programmes or individual modules in the field of artificial intelligence. On the other hand, universities are supported in the design of AI-supported learning and examination environments.