Technology can mitigate – reverse powers

Artificial intelligence (AI) can help authorities and organizations with security tasks (BOS) in their work and put the employees there at ease. This applies, among other things, to the evaluation of abuse data and recordings. The head of the Central Office of Information Technology in the Security Sector (ZITiS), Wilfried Karl, was convinced of this.

There are already machine learning systems in use. In light of the many challenges that BOS will have to face in the field of data, Karl said at the Day of Digital Police to Spiegel authorities that there is an urgent need as well. For example, the amounts of data to be analyzed are becoming larger and more diverse. In addition, there is often time pressure to investigate. Artificial intelligence can make work easier here. However, technology will not replace BOS employees for the foreseeable future. However, the personnel requirements profile will likely change, Carl predicted.

Maintaining digital sovereignty

The head of ZITiS called for Germany’s digital sovereignty to be preserved. Complete self-sufficiency is not required for this. However, the possibility of independent work is always needed, especially in terms of resources and technologies. In Karl’s view, it is important to buy AI not based on closed systems, but on modularity and open interfaces. However, artificial intelligence is not a panacea, especially since the term is sometimes used in an inflated way.

Algorithms are dynamic

Carl received encouragement for this statement from Dr. Martin Thüne of the Thuringian University of Applied Sciences of Public Administration. Also from his point of view, the term is sometimes used in a striking and inconsistent way, which leads to mistrust. However, it is clear that algorithm-based policing plays a central role in AI technology.

According to the scientist, administrators and users expect this to improve assessment processes through (micro) automation, rapid processing of text data (heterogeneous), images, video and audio as well as enabling faster, more efficient and more flexible analytics while ensuring trustworthiness and traceability of core processes. However, in Thun’s view, not all predictive police applications currently in use are actually artificial intelligence technology.