In the area of IT criminal law the effort required to secure evidence and analyze the data is increasing due to the constantly increasing amounts of data over the years despite the increased computing power. The use of appropriate forensic tools is therefore necessary. These technologies help forensic scientists obtain results in a short time that they would never have obtained if they had been examined manually. However, they should not allow themselves to be carried away by such results and make hasty and possibly incorrect conclusions. This phenomenon, known as “bias” is the systematic incorrect tendency in perceiving, remembering, thinking, and judging. The AI systems that have emerged recently could therefore represent an innovative method of supporting experts in arriving at results quickly on the one hand, and on the other hand avoiding bias through this “artificial intelligence”, and thus increasing the quality of the findings and reports. This work deals with the possibilities of using AI sensibly in expert reports and also shows the limitations using examples. Using “Google Bard”, it was researched that the generation of technical requirements, such as a test program or a data model with test data, delivers almost perfect results, which saves time for expert work. On the other hand, it can be shown that when questions are asked in an interdisciplinary area, such as drawing legally relevant conclusions, AI systems fail completely with incorrect results.