Government in digital dialogue: Using AI to create efficient and user-friendly administration

Artificial intelligence (AI) is increasingly changing the interaction between the state and its citizens and businesses. With the aim of creating an efficient, networked administration that is geared to the needs of people and businesses, public authorities are increasingly relying on automated processes, intelligent data analysis and personalised services. However, government services should not only be faster and easier to access digitally, but also trustworthy and efficient. But what legal, normative and target group-specific requirements are needed for AI-based interaction in administration? Particularly in the areas of automated administrative acts, digital approval procedures and AI-supported information processing, profound legal, ethical and social questions arise. Using examples, Plattform Lernende Systeme shows in a new white paper how government action can be made more efficient, accessible and user-friendly through AI technologies and presents options for successful implementation.

AI is considered a key technology for the digital transformation of public administration. While online forms and automated data transfer merely map standardised processes digitally, AI can process information in a context-sensitive manner, tap into complex data structures and provide individual support in real time. This creates entirely new forms of interaction that are particularly advantageous when procedures are complex. As a result, interaction between the state and its citizens and businesses becomes more efficient and, at the same time, more accessible and user-friendly. Using concrete examples from practice, such as the Hamburg chatbot ‘Frag den Michel’ (Ask Michel), the white paper demonstrates the potential of AI use in administration.

Analogue and digital access must be considered together

In the course of digital transformation, access to government services must also be linked back to the analogue world. Especially in the interaction between the state and citizens or businesses, analogue and digital access to administration must be considered together in order to ensure inclusion, participation and equal treatment in administrative actions. Older people and those with limited digital skills in particular need some form of human-centred support when dealing with AI systems. This could be provided, for example, through low-threshold advisory services, hybrid interfaces or so-called ‘AI guides’ who accompany the transition between analogue and digital interaction and build trust in new procedures.

At the same time, the specific requirements for interaction differ depending on the target group. For citizens, personal circumstances are often at the forefront, for example when applying for social benefits, residence rights or parental allowance. In such sensitive contexts, AI systems must pay particular attention to comprehensibility, accessibility and emotional dimensions. Where decisions have existential implications, low-threshold access to appeal procedures is just as essential as the possibility of contacting a human representative. Interaction with companies, on the other hand, often involves compliance with regulatory requirements, such as reporting obligations, approvals, tax matters or applications for subsidies. These processes are often highly standardised and therefore predestined for technical automation.

Responsible use of AI 

‘The use of AI in public administration is changing not only internal processes, but above all the interface between the state, society and the economy,’ explains Prof. Dr. Christoph Bieber, member of Plattform Lernende Systeme. ‘The responsible use of artificial intelligence in public administration therefore also requires joint action by the state, society and the economy. If AI systems are designed to be fair, transparent and participatory, trust in public institutions can be strengthened, contributing to a more effective and people-centred administration.'

The white paper lists key criteria for the responsible use of AI in a governmental context: AI systems should be transparent and explainable, and they should be controllable and questionable in interaction. Furthermore, when using AI systems in administration, it must be ensured that the requirements for transparency, control and legal protection are proportionate to the context of application. If the use of AI systems is also inclusive, user-oriented and characterised by a clear allocation of responsibilities, it can be ensured that AI systems are not only lawful but also socially acceptable.

About the whitepaper

The white paper ‘Der Staat im digitalen Dialog. KI in der Interaktion zwischen Staat, Bürgerinnen und Bürgern sowie Unternehmen’ was written by members of the ‘IT Security, Privacy, Law and Ethics’ and ‘Innovation, Business Models and Processes’ working groups of  Plattform Lernende Systeme. It is available for download free of charge.