AI for Public Administrations: Strengthening Sovereign And Effective Governance
Processes in public administrations can be long-winding and repetitive, yet they are essential for a well functioning state. Thus, it comes to no surprise that citizens are expecting their public administrations to keep up with the digitalisation of today’s world. When it comes to AI, several countries have developed strategies to integrate the technology into the public sector. Of course, public administration brings with it a number of considerations, especially in the realm of data security and privacy. Some of these can be mitigated by a sovereign European approach.
For Europe, the digitalisation of the public sector is especially important considering its aging population. Many public sector officials will retire in the next decade and there are fewer people to replace them. Considering this, employing AI can be a strong solution. Nonetheless, human oversight is still extremely important since outsourcing decision-making to AI can also do more harm than good.
How can AI support the Public Sector?
As we have seen through the adoption of LLMs in our daily lives, AI can shorten tasks that would otherwise take much longer. In comparison to older tools, it has the power to process natural language and recognise patterns in ways that can easily accompany human decision making. Possibly most importantly, AI tools that use natural language are more intuitive to use than any predecessor, even for less tech-savvy users. These advantages can and should be leveraged by the public sector.
Due to the nature of this sector, putting sensitive citizen data into e.g. ChatGPT is not an option since it leaves it in the hands of non-European companies. Not only that, but there have also been cases in which AI assistants have been influenced to leak sensitive data that had been fed to them by users. Furthermore, there are a variety of other AI powered solutions available that are not LLM-based chatbots which is why this article will offer a broader look at how AI can be used in the public sector.
Supporting Internal Processes
Increasing the efficiency of processes is one of the best examples of how AI can strengthen the public sector. This includes translation/transcription, document summary, and report generation used in Germany, as well as Optical Character Recognition to convert handwritten text into scanned forms and Robotic Process Automation for repetitive tasks employed in parts of the Canadian public sector. These types of AI-powered solutions are especially useful because they speed up internal processes which citizens ultimately profit from.
Transforming citizen-oriented Services
The primary goal of a public administration should be the delivery of services for its citizens. While AI chatbots are an old hat these days, countries like Italy have been using AI to match job-seekers with job openings, Canada is using natural language processing to make employment insurance processes more efficient and the European Commission has enhanced its funding and tender portal with AI for a better user experience. In the realm of fraud detection, the US has been using AI to uncover corruption while Canada is using it to check submitted documents for fraud. All of these examples show how AI can be used to improve citizens’ lives and aid the overall goal of offering government services online, thus making them more accessible.
Improving the Policymaking Process through data-driven Insights
In order for policies to reach their intended targets, they are best based on data. Here, AI can support through keyword and sentiment analysis and monitoring as well as by giving overviews of legislative texts and existing policies. The UK and the European Commission for example, use AI to analyse the feedback they receive in their public consultations while Germany analyses gamma radiation levels and the sale of endangered species online via AI. This shows how analysis with AI can be used at different stages of the policymaking process. From the early formulation stages through to the implementation of policies.
Facing Challenges head on
As with any technological change, there are a few challenges that need to be addressed in order to achieve a fully functioning system. Facing these challenges head on might require some consideration but will facilitate operations massively once in place. Thus, it is worth taking a look at what can be done to smoothen the way for digitalisation.
Interoperability is Key
The previous examples have shown how public administrations around the world have started to integrate AI into their processes. They have also highlighted that there are still large gaps in the use of AI in the public sector, even in countries that are starting to embrace the change. These AI projects feel more like lighthouses than big interconnected cities. For example, while the German employment agency might have embraced AI support for its internal services and the country’s social insurances are now updating to the cloud, only partially EU-based which might become an issue with the new Trump administration in office, there are still many agencies that have not done so. This is especially concerning considering that digitalisation works best when its components are interoperable and are not just adopted in one part of the administration.
The Power of Public Procurement
Procurement can be considered from two perspectives, on the one hand, the difficulty of procuring trustworthy AI for public sector officials who might not have the background knowledge to judge the product. Here, a database of trustworthy AI like the one in the UK can provide guidance. On the other hand, public administrations can also support the development of good and trustworthy AI through partnerships and procurement. They can offer reliable contracts and in the case of public-private partnerships, the long-term commitment that the design and training of an AI needs.
Upskilling and Reskilling
AI literacy is one of the most important aspects when it comes to the successful integration of this technology in the public sector. It is thus crucial to invest in the training of officials to ensure the best possible outcomes. At the same time, the introduction of AI might lead some officials to feel like they will be replaced. To ensure that does not happen, public administrations should draw up strategies to reskill and possibly upskill those who might have previously taken on jobs that are now supported by AI. In Germany, the GovTech campus has been offering learning workshops which were attended by over 2000 public servants in 2024.
AI Act - Burden or Opportunity?
The AI Act of the European Union will have a major impact on how and which AI will be trained in the EU in the upcoming years. Due to the sensitivity of the data, many of the AI applications for the public sector fall under the category of high-risk or unacceptable risk such as assessing eligibility to benefits and services and social scoring respectively. This regulation should not discourage the use of AI in the public sector, it merely shows how important the choice of the right AI tool is.
Sovereign European Administrations for a sovereign Europe
As we have also seen through the emergence of LLMs in our lives, AI can provide false information, hallucinate or spark discussions on privacy laws. Placing a focus on well-designed and sovereign AIs for the public sector is key. Sovereign public administrations are one of the most essential aspects of European tech sovereignty. How can sensitive citizen data be protected if it is fed into databases and stored in clouds owned by big tech companies making them reliant on the security but also the goodwill of those players? It is thus essential to employ AI that can be trusted to adhere to European standards of privacy and security and where data can even be stored onsite.
How to move forward?
- Leverage best practices from other countries to strengthen public administrations on all levels - from the city to the federal level.
- Use procurement and partnerships to strengthen the training of sovereign European AI for the public sector.
- Invest in AI literacy and potential reskilling for public sector employees to ensure the safe use of AI.
- Invest in regional cloud projects to retain European data in Europe instead of storing it in third countries.
- Promote the digital offers that are already available to the public, so citizens can appreciate that progress is being made.