Artificial intelligence applications are widely used. They not only offer us options, but often decide autonomously by presenting us with already optimized "solutions." Why didn't I get this mortgage? Why am I being included in the target group for this ad?  "The answers to questions such as this are simply not available to us as laypeople, because some factor triggers a chain reaction and the algorithm calculates a result based on that," says Titus Udrea, who researches this topic at the Institute of Technology Assessment (ITA) at the Austrian Academy of Sciences.

"Because AI systems are highly complex, we have only a very limited understanding of how they work and what factors ultimately influence the results and how. We therefore accept these gray areas as given. But how then can we distinguish iwhen and in what context AI systems are harmless or dangerous?", stresses Walter Peissl, deputy head of the ITA and coordinator of this study.

This short study, which the ITA is carrying out in cooperation with the Austrian Chamber of Labor, examines the extent to which AI systems can be designed to be comprehensible and transparent for consumers. It was triggered by the EU Commission's proposal for an EU-wide AI law, which is to provide a legal framework for the development and use of trustworthy AI systems.