Digitisation promises an increase in efficiency and effectiveness in almost all areas of state responsibility and administrative action. The proclaimed objective of digital administration is to improve performance and service, but at the same time to reduce expenditure and unnecessary bureaucracy. Last but not least, the potential of digitisation can be found in an increased transparency in terms of availability and accessibility of information (open data, open government) and stronger participation of citizens in processes of political opinion-shaping and decision-making. By means of new technologies, transformations of entire authorities and completely new administrative products are conceivable. However, the fact that Germany shows a structural deficit regarding the digitisation of its administration is raised as a problem in political discussions. At the same time, various activities have been initiated in recent years to improve the situation. Last but not least, the Online Access Act (“OZG”) was passed in 2017 with the aim of making all public administration services digitally accessible to citizens and businesses via online portals by 2022.

In the current professional public debates, reference is made mainly to the potentials of systems using artificial intelligence (AI) and distributed ledger technology (DLT) with regard to the field of public administration. AI technologies are associated with autonomous and data-based decision-making processes, the diverse application scenarios of which can generally result in an increase of effectiveness, quality and security of administrative processes. The same applies to DLT and blockchain, which – by means of distributed data structures – could make it possible, for example, to efficiently and securely design transaction and process management in connection with registers and land registers and to automate them across authorities.

This TA study is based on a prospective analysis of the fields of technology AI and DLT as well as their innovation potential with regard to the further development of public administration in Germany. In addition, the study includes relevant European and international practical examples of the digitisation of administrative processes the technological basis of which are AI- or DLT-based developments. Based on these findings, an in-depth analysis of the fundamental challenges that public administration faces when using AI and DLT is carried out.

Key results of the project

The analyses on the status quo of the use of AI and DLT in Germany and abroad show that both technologies are being used in more and more fields of application at all levels of public administration. Thus, in Germany, various AI applications can be found in public administration at the federal, Länder and municipal level. Nevertheless, it is evident that the number of piloted or implemented AI applications in public administration is still low compared to the private sector – and the use of AI in this context is still rather a niche topic. Possible reasons for this are a lack of know-how, but also a certain resistance due to possible job changes and responsibility problems with AI-based decisions as well as legal obstacles (concerning data protection). Many of the DLT projects identified in Germany’s public administration are only in the planning phase or at an early testing stage as pilot projects. In this context, a survey of heads of public authorities shows that the knowledge and skills regarding DLT in Germany’s public administration are not very well developed so far.

An international comparison with regard to the use of AI and DLT in public administration shows that in many countries, a number of digitisation projects have already been transferred into the regular operations of the administrations. In addition to individual applications, such as automated traffic control or verification of documents and certificates, comprehensive digital infrastructures have been introduced as a basis for various administrative services (registering one’s place of residence, applying for support services such as child benefit).

With the advancement of technologies and the development, piloting and adoption of AI and DLT applications in regular operations, public administration in Germany is facing not only the opportunities of tapping their application and innovation potential, but also new challenges regarding their use.

Challenges regarding AI applications in public administration include i. a. the quality of the underlying AI model as well as the availability and quality of the data base used. Thus, for example, a lack of understanding of underlying models and procedures might lead to misinterpretation of the results within the framework of administrative action or to the results of AI applications not being critically questioned. With regard to the quality of the data base, there is a risk that any structural biases in the underlying data base will be perpetuated within the AI-based classification, forecasting or recommendation decisions. However, for tapping the innovation potential of AI, both explicit expertise and implicit empiric knowledge are among the areas that have so far been difficult to build up within public administration itself and represent a prerequisite for the use of AI applications in regular operations.

Some of the challenges with regard to the introduction of DLT applications in public administration are the complexity of this technology, increasing and distributed data volumes and thus also growing requirements regarding data storage. Moreover, DLT applications must be integrated into existing administrative processes and IT systems – an aspect to be taken into account for the development and implementation of procedural security concepts of the respective DLT applications. In addition, creating and operating DLT applications requires specialists from data science, cryptology and computer science. Furthermore, DLT applications require technical configurations to be tested with regard to their energy expenditure and to be evaluated in terms of energy and resource consumption.

Fields of action derived from the results can be found primarily in the areas of management and coordination of digitisation projects, competence development and knowledge management in public administrations as well as in the areas of adapting regulatory framework conditions and designing cross-authority applications.