AI – a game changer for the state of the future?

Indeed: The structures of many democracies that have evolved over decades too often get lost in complexity, an opaque landscape of actors, lengthy bureaucratic decision-making processes and political measures developed without a valid data basis, whose effects sometimes come to nothing despite costing enormous sums running into billions.

‍

Modern state structures are the foundation of capable democracies. How important this is becomes particularly evident in times of crisis. Whether registering hundreds of thousands of refugees, disbursing economic aid during pandemics (keyword: coronavirus aid) or paying energy allowances to students – with an efficient state and administrative structure that works seamlessly across the levels of municipalities and states up to the federal level, many crisis responses and political action in general could be faster, more effective and more cost-efficient. 

‍

This affects trust in our political system. Germany as an example: as a study by the Konrad-Adenauer-Stiftung shows, only just over one in three Germans (36 per cent) are satisfied with the actual functioning of democracy in Germany.

‍

As a result, calls for comprehensive modernisation of public administration are growing louder across all political camps. “Administrative modernisation” is suddenly no longer a niche topic for specialist legal experts in politics, but is moving to the centre of the political agenda. The question of the state’s capacity to act and organisational model in times of rapid progress, especially in new digital technologies, will shape future election campaigns. 

‍

Artificial intelligence (AI) can become a crucial tool for modernising the state. AI has developed into one of the central drivers of digitalisation and is of enormous importance not only for the economy and society – but also for our democracy.

‍

In fact, AI applications have long been tested in public administrations and are creating real benefits, even if this often goes unnoticed by the public. Public administration in Germany in particular is an active tester of AI: from digital inclusion through AI-developed plain language for administrative information, and the preparation of parliamentary documents, to the use of large language models (LLMs), for example for text summarisation or support in drafting memoranda and legal submissions – to name just a few examples. 

‍

Nor will it be limited to supposedly simple use cases. 

‍

From the use of AI in tax and financial authorities (“Financial Intelligence”) to process mining and the preparation of untapped data and information in public administration organisations – in future, AI will be used on a far greater scale as a key technology by governments and public sector organisations. 

‍

In doing so, governments must design and implement AI applications in such a way that they are secure, contribute to supporting democracy and process data (of both the government and citizens) transparently, securely and reliably. This is because the use of AI in the public sector requires citizens’ trust in the new technologies just as much as it requires public administration employees’ trust in the added value and benefits of the technology for their everyday administrative work. 

‍

As part of the European AI Act, the world’s first comprehensive regulatory framework for AI was developed. If it proves successful in practice, the European law will provide the decisive framework for the use of artificial intelligence in Europe, promoting innovation while strengthening trust in AI. An initial success with a global role-model character that individual national regulations can scarcely achieve.

‍

But even if the potential for using AI has been recognised and an initial AI regulatory framework has been developed, we should not expect miracles, especially in the early stages. Politicians in particular overestimate the potential applications of AI in the short term, while underestimating its long-term effects. If inflated expectations are disappointed, the urgently needed structural changes for the widespread use of AI may not be addressed. That is why realistic expectation management is important.

‍

This is also clear: AI applications are not the ultimate panacea. Their use alone will neither make the state more capable of acting nor significantly reduce dissatisfaction with bureaucracy and administrative structures. It will not be possible without fundamentally restructuring our public administration apparatus and taking long-term measures such as investing in the (digital) infrastructures of the future – above all, the cloud. 

‍

The principle here is to finally make the state and public administration platform-ready. Digital platforms are fundamental to providing digital solutions based on shared standards and development principles. So far, public-sector actors have paid little attention to the proactive use of platform principles by the state; a seemingly unmanageable number of public and semi-public actors still too often work in parallel structures, without coordination or collaboration across different boundaries. 

‍

Changing this is the task of consistent, determined political leadership. Far too often, attempts are still made to build “the new” on the political operating system of the past. Major success has failed to materialise despite billions being spent on modernising public administration. And that is hardly surprising, because the success factors of innovation and the future no more follow the administrative boundaries of federal states than they do the politically carefully calibrated, closed ecosystems of the state’s “IT age”, which ended long ago. 

‍

More than ever, the state and public administration must build open ecosystems in order to make democracy technologically capable. Tech companies, start-ups and research players in emerging fields such as machine learning and data science have the expertise and capabilities that the state must make use of. Public-sector actors will find it difficult to keep pace, particularly in the AI environment. However, this must not lead to one-sided dependency. To be able to operate on an equal footing, public administrations must build the relevant capabilities to effectively establish, manage and assess partnerships between the state and AI companies. Only in this way can it be ensured that public institutions involve companies’ AI expertise in a targeted, application-specific manner and are able to assess the design of AI applications. 

‍

As with every new technology, AI offers far-reaching opportunities if we accept the risks, do not exaggerate them – and contain them as effectively as possible. Constructive, ambitious engagement with the possibilities that AI offers us is certainly worthwhile. After all, the intelligent use of AI can make a significant contribution to enabling a more capable and technologically capable state; and, by improving the quality of public services, to strengthening trust in democracy once again.