Unfold your career

Kostenlos abonnieren

Werden Sie regelmäßig per E-Mail über neue Ausgaben der campuls informiert. Sie können Ihr kostenloses Abo jederzeit einfach online über den Abmeldelink im Newsletter kündigen.

Weitere Infos zu Datenschutz & Widerrufsrecht finden Sie hier.

More Than Just Prompting: What Skills Do We Still Need to Have Ourselves When AI Can Do (Almost) Everything?

Hof University of Applied Sciences is participating in the Bavaria-wide “AI Learning Lab” project. Together with students and faculty, it is exploring how people can continue to learn and work independently, creatively, and with sound judgment even in the age of artificial intelligence.

In the photo, from left: Prof. Alexander Forst, students Andreas Irrgang and Jana Brücher, and Anke Kaluza (Center for Teaching and Learning at Hof University of Applied Sciences); Photo: Hof University of Applied Sciences;

Developing a new bicycle sounds, at first glance, like a task where artificial intelligence can fully showcase its strengths. A brief assignment is all it takes—and the AI immediately delivers a design, develops variations, and makes suggestions for materials or functions. What might once have taken hours or days is now completed in a matter of seconds. But this is precisely where the real question arises: What was learned in the process?

Hof University of Applied Sciences is addressing this question as part of the Bavaria-wide AI Learning Lab. Six teams from various universities are each working on their own task and investigating how artificial intelligence can be used in a meaningful way that promotes learning in both study and teaching. What makes this project unique—and what should ideally become the norm—is that students are actively involved in addressing the questions and developing solutions from the very beginning. The Hof team consists of Prof. Alexander Forst, Anke Kaluza from the university’s Center for Didactics, and students Jana Brücher and Andreas Irrgang. Together, they are tackling the challenge “More Than Just Prompting—Learning, Thinking, and Deciding with AI.”

Not Just Faster—But Still Independent

The focus is not on how to use a specific AI tool as skillfully as possible. After all, individual applications can change within a few months or be replaced by new systems.

Of course, working with specific AI tools is part of it. But what’s crucial is what students need to be able to do, regardless of the tool in question.”

Prof. Alexander Forst.

This includes, for example, the ability to ask good questions, critically evaluate AI results, and recognize when a result is scientifically sound. Or to put it another way: The crucial skill isn’t simply getting an AI to provide an answer. Rather, the question is: How do we ensure that people develop their own perspectives and decision-making criteria despite the possibilities offered by AI? And how do we preserve space for curiosity, detours, and non-goal-oriented thinking when AI can very quickly lead us to a seemingly appropriate solution?

Who decides when the answer is already there?

Of course, AI can provide information, ideas, or proposed solutions in a very short time. But not every answer is automatically relevant or factually correct. People must continue to assess which information is relevant to their own question, which suggestions make sense, and where closer examination is necessary.

We want to investigate how learning processes must be designed so that AI provides support without replacing our own thinking and decision-making. This also involves tolerating uncertainty, exploring different approaches, and not always seeking the most efficient answer right away.”

Anke Kaluza

This shifts the role of humans: away from the mere production of content, toward its classification, evaluation, and further development. The same likely applies to learning: While those who let AI generate a solution for them can arrive at a result quickly, Whether genuine understanding has actually been achieved depends on how the learning process is structured. This is precisely where the Hofer Challenge, as part of the Learning Lab, comes in. The team aims to systematically investigate where AI enhances learning and thinking processes—and where it might cut them short too soon.

The Bicycle as the Perfect Analogy

This can be illustrated particularly clearly using the example of product development: An AI can generate a design for a new bicycle at the push of a button. But instead of considering the task complete at that point, students could first develop the key questions themselves: What must a bicycle be capable of in five years? How can it support the mobility needs of an aging society? What sustainable production options are available? Only then does AI come into play. It can further develop ideas, generate variations, or challenge assumptions. But humans continue to set the direction.


“What remains crucial is which questions we ask, which paths we pursue, and how we evaluate the results.”

Jana Brücher, student

This transforms AI from a mere solution generator into a thinking partner. It can open up possibilities and provide food for thought—but the decision about which of these to pursue remains with humans.

“This is particularly important when it comes to innovation. Not every good idea comes about directly. Creativity also requires detours, unconventional thinking, and the possibility that something might not work at first.”

Prof. Alexander Forst

The Temptation of the First Plausible Answer

At the same time, those involved certainly see risks. AI can tempt us to settle very quickly for the first plausible result. If the machine delivers a seemingly suitable solution, we may become less willing to ask further questions, develop alternatives, or truly delve deeply into a problem. Decisions, too, could increasingly be delegated to AI. This threatens to erode something crucial for learning and personal development: one’s own point of view and one’s own criteria for making good decisions. Anke Kaluza makes it clear, however:

“Our goal is not to keep AI out of the learning process. On the contrary: We want to investigate how learning processes must be designed so that AI provides support without replacing independent thinking and decision-making. This also includes tolerating uncertainty, exploring different approaches, and not always seeking the most efficient answer right away.”

Anke Kaluza

Student Andreas Irrgang therefore sees actively engaging with AI as a particularly important task:

“We must learn to use AI sensibly without handing over responsibility for our decisions to it.”

Andreas Irrgang, student

The Hof team therefore does not want to limit the use of AI in academic studies. Rather, the question is how learning processes must be designed so that AI provides support without replacing students’ own thinking and decision-making.

Four Questions for the Future of Higher Education

Anke Kaluza sums it up: “The AI Learning Lab Challenge raises four key questions: How do students remain creative when AI already provides numerous ideas? How do they recognize which AI outputs are actually viable? How do they keep track of the ever-growing number of AI tools? And: How do they continue to learn independently when answers are available at any time?”

Ultimately, however, this raises a very fundamental question: What skills do people actually still need in a world where knowledge, information, and proposed solutions are available at any time and in nearly unlimited quantities? Traditional “future skills” such as critical thinking, creativity, self-directed learning, and decision-making skills are thus becoming increasingly important. Added to this is the ability to continually adapt one’s own learning to new technological developments.

Students are not test subjects, but co-creators

This is why Jana Brücher and Andreas Irrgang play a special role. Both contribute their own experiences with AI as well as the perspective gained from the master’s program in Artificial Intelligence-Aided Mobility Design. They are not merely test subjects for a concept developed by faculty members. They collaborate on the research questions and solutions from the very beginning.

For Prof. Alexander Forst, this is a key principle:

“We should design good teaching not just for students, but together with them.”

Prof. Alexander Forst

The experiences of the students in Hof demonstrate how AI is actually transforming learning and what kind of support will be needed in the future. At the same time, students themselves can help shape how roles will be distributed in the future: What will AI take on? What will remain the responsibility of faculty? And what skills must students develop so they can use AI confidently without relinquishing their own judgment to it?

From Experiment to Concrete Learning Program

The AI Learning Lab project is scheduled to last about one year. Following the three-day kickoff in September in Nuremberg, the project will now proceed with work phases within the individual teams, as well as joint in-person meetings in December 2026 and in April and June 2027. The results are scheduled to be publicly presented at a closing event in September 2027. Work has already begun for the Hof team. At the kickoff in Nuremberg, the original research question was further developed. Among other things, the participants examined where friction arises in the interaction between humans and AI, what potential this holds, and which questions need to be pursued further.

In the coming months, this will be developed into a concrete learning format or support program, which will then be tested. The central question remains surprisingly simple:

If AI can provide us with answers faster and faster—how do we ensure that we don’t lose the ability to ask the right questions ourselves?


Rainer Krauß
published: 19.11.2026

Weitere Themen