How can AI-supported tools improve teaching and learning at universities without losing the campus as a social learning space? Representatives from science, politics and university practice discussed this topic at the TUM Campus Garching on 17 November 2025, focusing on AI tools and current approaches to digital teaching as well as the initial results of the bidt research project “Effectiveness of Generative AI Tutors in Higher Education (AIffectiveness)”.
Bavaria’s Minister of Science Markus Blume, MdL, took the event as an opportunity to emphasise the strategic importance of AI for university teaching. Bavarian universities should play an active role in shaping this development.
AI in the lecture hall?
In his speech, Science Minister Markus Blume recalled earlier digital upheavals – from the first internet access at universities to the introduction of search engines – and categorised AI as the next stage of this development. For him, it is clear that AI will play a permanent role in university teaching:
When it comes to innovative teaching, AI is not a 'nice-to-have' but a 'must-do'. AI is here to stay. On the contrary, it will develop explosively. Our universities are the perfect place to try out new things and evaluate them at the same time. We must use AI in such a way that we can develop our human talents even better.
Markus Blume, MdL
At the same time, Blume focussed on responsibility and framework conditions. AI should provide support, but not replace human expertise. It is crucial to actively shape technologies and set clear framework conditions, especially with regard to examination formats and study organisation:
The same applies to teaching and learning formats: the greatest danger with AI is not being involved. From AltaVista to Google to ChatGPT – the crucial thing was and is that we use technologies as support systems and actively set guidelines. This also includes adapting our examination culture. With the amendment to our Bavarian Higher Education Innovation Act, we will ensure that a general ban on artificial intelligence in examination regulations does not make sense.
Markus Blume, MdL
AI in teaching: examples from practice
OneTutor is an AI-based teaching and learning tool that is integrated directly into lectures. Students can use it to ask questions about the lecture material, have content explained to them and complete customised quizzes for revision. The idea for the AI-supported learning assistant was developed as part of a student project at the chair of Professor Alexander Pretschner at TUM, which has since resulted in the founding of a start-up. OneTutor is now used at around 30 universities in Germany and Austria, with over 21,000 active users in more than 620 lectures.
From the project team’s perspective, the focus is not on improving grades “at the touch of a button”, but on making learning more accessible overall: According to their observations, those who actively use the tool engage with the material more frequently over the course of the semester. Initial analyses indicate that students using OneTutor fail exams less often. However, the quality of the teaching itself remains crucial for the team behind the application: Good AI tutors can only be as good as the courses on which they are based.
In addition to OneTutor, the university assistance system HAnS and programmes offered by the Virtual University of Bavaria (vhb) also showed how AI is already being used in teaching.
HAnS, developed as part of a Bavarian funding project and presented by Professor Tobias Bocklet (Technische Hochschule Nürnberg Georg Simon Ohm), automatically links transcribed teaching videos with chapters, supplementary materials and an integrated search and chat function. This allows students to jump directly to relevant passages and rework content across different media. The project is funded by the Federal Ministry of Research, Technology and Space (BMFTR) as part of the federal-state initiative “AI in Higher Education” and the “Digital Higher Education” funding programme.
On behalf of the vhb, Alexander von Stetten explained how AI elements are gradually being incorporated into existing online courses – from assistance functions within learning environments to specially funded units for building AI skills.
Together, these examples emphasise that different ways of meaningfully embedding AI in digital and hybrid teaching and learning formats are being tested in Bavaria.
Accompanying research at bidt: How effective are AI tutors?
In parallel to the practical use of OneTutor, the bidt is investigating how such systems influence learning behaviour and learning success in the project “Effectiveness of Generative AI Tutors in Higher Education (AIffectiveness)”. To this end, surveys of students and lecturers were combined with usage data from the courses. Ten Bavarian universities are involved, including universities, universities of applied sciences and the Bavarian Virtual University (vhb). In the first semester of the accompanying research, results from 55 courses were analysed.
Initial key findings:
- According to the “bidt Digital Barometer 2025”, 88 per cent of pupils and students already use generative AI. AI has thus arrived in everyday education.
- The majority of lecturers rate OneTutor as a useful addition to teaching. Around 90 per cent agree with the statement that the tool complements existing learning methods in a helpful way.
- Students who use OneTutor see it as one of the most important learning options for exam preparation and report that it helps them to understand difficult content and stay in the learning rhythm.
- Students who do not use OneTutor do so mainly because of personal learning preferences, for example because they prefer to work with their own notes or to exchange ideas with others. Quality concerns about the tool play a much smaller role.
The initial analyses thus show a differentiated picture: Generative AI is already present in university teaching. Whether it is used depends heavily on the individual learning style. For the design of AI-supported programmes, this means that flexible scenarios are needed that reach different student groups.
AI has the potential to make teaching and learning more individualised and personalised through constant feedback between lecturers, students and the new tools. We now need to understand which factors lead to measurably better learning outcomes – because if AI tools externalise thinking, this is of no help to anyone, especially in higher education. Our initial experiences with OneTutor have made us very confident!
Prof. Dr. Alexander Pretschner To the profile
To the research project
Student perspective: AI as an opportunity – but not without people
Three computer science students presented their views on AI in teaching in short statements:
- AI has long been an integral part of everyday life, studies and the future world of work for Generation Z. Universities are the right place to test and critically scrutinise specialised AI tools.
- At the same time, educational processes should not be left to companies such as OpenAI alone. Universities have a responsibility to provide fair and free AI offerings and to actively participate in their further development.
- Even in a future with very powerful AI systems, professors, tutors and especially fellow students cannot be replaced: Motivation, role models and shared learning processes would remain tied to people.
Subjects such as programming in particular show how “double-edged” AI can be: copy-and-paste solutions without personal reflection are tempting, but not sustainable. This is where a well-designed AI tutor comes in and opens up new opportunities to specifically clarify questions of understanding, reflect on ways of thinking and thus promote sustainable learning.
Panel discussion: What will the university teaching of tomorrow look like?
In a concluding panel discussion, Minister of State Markus Blume, Professor Tobias Bocklet (Nuremberg Institute of Technology), Alexander von Stetten (vhb) and Professor Alexander Pretschner (TUM/bidt) discussed the future role of AI in teaching. It became clear:
- Common basic AI equipment: Blume campaigned for a common understanding and “basic AI equipment” for all Bavarian universities. Solutions are needed that are scalable and at the same time do justice to the different subject cultures.
- Classroom teaching remains central: The panellists agreed that AI would not replace classroom teaching. Blume emphasised that a lecture must be a “happening” – a privileged space in which charisma, exchange and joint reflection come together, even in times of AI.
- The role of lecturers is changing: Von Stetten and Bocklet emphasised that the role of lecturers is continuing to shift towards support and moderation. AI can support knowledge transfer, but practising, applying and discussing together remains orientated towards smaller face-to-face formats.
- “Campus exodus” as a challenge – and as an incentive: Pretschner pointed out that a trend towards “campus exodus” already exists today and is not triggered by AI alone. Blume also saw this as an incentive to further develop teaching in a targeted manner and thus clearly stand out from purely digital standard offerings.
The panellists were unanimous in their perspective on AI in universities:
AI-supported tools such as OneTutor have the potential to make learning more individualised and accessible. However, whether they realise this added value depends crucially on how they are designed, where they are used and under what conditions they are combined with face-to-face teaching, human support and institutional strategies.
The bidt is accompanying this process with the research project “Effectiveness of Generative AI Tutors in Higher Education (AIffectiveness)” and thus providing a basis for evidence-based decisions in universities and politics.
Further links
to the research project
