Trainees trust AI more than their teachers

AI has long been part of everyday life in company training: 85 percent of trainees already use tools such as ChatGPT, Claude or Gemini in the training context. This is shown by the study “Trainee Recruiting Trends 2026” by the software and test provider U-Form in collaboration with the Koblenz University of Applied Sciences. According to the study authors, the study is the largest survey on dual training in Germany to date; its results are available to the human resources industry in advance. Over 11,500 people took part, including almost 9,500 school students, trainees and dual students as well as around 2,000 people responsible for training.

AI versus humans: teachers lose the comparison

AI seems to improve the learning process of young talent and performs well as a teacher function compared to trainers. 53 percent of the trainees surveyed think that AI can explain topics more often or sometimes better than their own company trainers. When looking at vocational school teachers, this figure rises to 74 percent. University teachers perform worst: 80 percent of dual students say that the AI ​​explains things better than their professors.

The ranking is revealing. In-company trainers come out best because their strength lies in practical instruction, an area in which AI tools have their limits, according to the study. Vocational school teachers, on the other hand, are traditionally responsible for imparting theoretical knowledge. This is exactly what generative AI is particularly good at: explaining content, providing examples, adjusting difficulty levels, and answering questions immediately. Those responsible for training themselves also confirm this: 62 percent report that trainees prefer to use AI when they have questions about training rather than contact their trainers.

Trainees are self-confident when it comes to AI skills

But in order to use AI as a teacher, you need AI skills. 85 percent of trainees rate themselves as “very fit” or “somewhat fit” when it comes to AI skills. And this despite the fact that they have largely taught themselves how to use AI.

Only 15 percent of trainees have ever taken part in further training in AI; most have acquired their knowledge through daily use and trial and error. Important: The study measures self-assessment, not actual competence. Whether trainees can critically classify AI output and recognize errors and hallucinations remains an open question.

In any case, companies are doing little to help trainees on this topic. 49 percent of training companies do not provide any learning opportunities on the subject of AI for their trainees. Only 40 percent of trainees believe that they can do without such offers. The “Trainee Recruiting Trends 2024” already showed that at that time only ten percent of training companies taught AI content in training. Little has moved since then.

Maybe because the trainers themselves first have to become AI-fit. So far, 39 percent of trainers have taken part in further AI training, more than twice as many as trainees. 54 percent of those responsible for training describe themselves as AI-fit.

AI use: What the 2026 vocational training report says

The “Vocational Training Report 2026” from the Federal Institute for Vocational Training (BIBB), published at the end of April, also shows that the gap between trainee practice and company offerings is not a marginal phenomenon. He dedicates a separate chapter to AI and provides figures that put the mood of the trainee study into a larger context.

According to the vocational training report, only around ten percent of employees at the skilled level use AI, while at the expert level it is around 37 percent. So AI is not yet a reality for skilled workers in the world of work. The BIBB states that it would be too short-sighted to conclude from this that the work tasks of skilled workers are not affected by AI. AI is also changing individual activities, text-related, organizational and information processing tasks at the specialist level. The dual training prepares you for a working world in which AI competence is needed at all levels. Most of the time it is not communicated.

Another finding of the vocational training report: In digitalized working environments, trainers increasingly no longer have the necessary knowledge, skills and experience advantage over learners.

Initial empirical findings cited in the report also show: Vocational schools and further education institutes use AI more intensively than training companies or chambers. Those responsible for training describe AI as helpful for creating materials and more individual learning processes, but express reservations about the learners’ critical thinking and ability to reflect.

What would be asked now

The BIBB report shows that employees in companies that are heavily committed to further training are significantly more likely to use AI. The institute warns: “If the participation in further training of skilled workers is significantly lower than that of employees at expert and specialist level, the risk increases that AI potential at the skilled worker level will remain unused.”

Sina Sommer, a consultant who designs learning and change formats, shows what this can look like in practice. She reports on Linkedin about a discussion with trainers in which there was a call to simply ban AI. Their counterargument: “Banning doesn’t solve the problem, it just postpones it.” Instead, she advocates a different task design: taking AI output as starting material, classifying it critically, questioning it and developing it further. Not “Let the AI ​​do it,” but “What’s wrong here, what’s missing, what would you decide differently?” That is the competence that counts in training.

Sven Frost is responsible for HR tech, which includes the areas of digitalization, HR software, time and access, SAP and outsourcing. He also writes about recruiting and employer branding. He continues to be responsible for the editorial planning of various special human resources publications.