Artificial intelligence has in a few years crept into most workplaces in Germany, but the great productivity revolution is yet to come. This is the main conclusion of a new, large survey among nearly 10,000 employees, which shows that almost two out of three use AI tools in their job. Nevertheless, only about one in five use them often. AI has thus become commonplace, but mostly as a tool that is occasionally brought out, not as something that is built into work processes.
The survey was prepared by a German research group with participation from, among others, the ifo Institute in Munich, IAB in Nuremberg, ZEW in Mannheim, and the Federal Institute for Occupational Safety and Health BAuA.
According to the researchers, the explanation may be that AI is spreading along two tracks. One is the informal track, where employees themselves start using tools like text-based generative AI because it is cheap, accessible, and requires few barriers. This accelerates the spread, but the use easily becomes sporadic and detached from the organisation's way of working. The other track is the formal rollout, where the employer introduces AI as part of the work. Here, AI is much more often something that is used regularly, and it is typically accompanied by more training and clearer perceived benefits in everyday life, both in speed, quality, and quantity.
Unevenly distributed
But there is a downside. When AI is formally introduced, it is also associated with more AI-based management, where systems can be used to allocate tasks, manage time, or assess performance to a greater extent. At the same time, the study points out that the spread of AI is uneven. It is especially highly educated individuals and employees in more complex jobs who adopt the technology, regardless of whether it comes from the bottom up or the top down. In other words, employers do not automatically get more people on board by rolling out AI.
The conclusion is that Germany has already experienced a broad, informal AI wave. But if AI is to truly boost productivity, it requires companies to invest in training, integration, and clear frameworks, without it becoming a new engine for inequality and control.
