Openness of administration: no more empty promises thanks to AI technology
Openness of Administration has been embedded in the Belgian Constitution for at least thirty years. Article 32 states that every citizen has the right to access administrative documents and to obtain an excerpt of these. After thirty years and thanks to growing digitalisation, you would think that this must be matter-of-course among our city councils, government agencies and other municipal bodies. Regretfully, there is still a lot to be done.
A bureaucratic labyrinth
Am I allowed to park my bike directly in front of my house? If there are roadworks blocking the way, how can I drive to work? What do I need to apply for if I want to organise a festival in the park? Do people need permission to fly drones in my municipality?
These are all commonly asked questions concerning everyday situations, but to which we can’t find an immediate answer on conventional search engines such as Google. Although the Law of 11 April 1994 on Openness of Administration states that you have the right to request (and obtain) information from municipal bodies, this process is often laborious, time-consuming and frustrating – not only to you, but also to the administrative staff who need to search for this information in ever-growing stacks of archives. It is literally a bureaucratic labyrinth through which it is impossible to find your way without an experienced guide.
Openness of administration is not only paramount to promoting transparency; it also helps citizens quickly find answers to specific questions.
The problem lies in the limitations of conventional search engines and chatbots. Despite their computational capabilities, these machines cannot think like humans. Administrative documents were not written to provide an answer to a specific question, asked by a human. A machine scans questions on keywords, but doesn’t understand the broader context, which means that it will never produce perfectly accurate results. If you were to enter the following text: ‘What do I need to apply for if I want to organise a festival in the park?’, you probably won’t get any information about permits, because you have not used this keyword in your query.
Serving citizens thanks to AI
Openness of administration may be embedded into the law, but because it is so cumbersome, it remains an empty promise. The solution to accelerating the digital transformation among municipal bodies can be found in AI tools offering text analysis, like the ones we build at Nalantis. They bridge the gap between conventional indexed search methods and advanced machine learning, which does understand the context surrounding your question. AI links the words ‘apply for’ to ‘permits’ through concept nets to achieve a more targeted result. Software developed with this technology can perform more targeted searches and provide you with more precise answers much faster.
For citizens and administrative staff
AI search engines are not only helpful to citizens, but also to the administrative staff working at municipal bodies. They can now respond much faster to questions posed by citizens, whether by email, by telephone, or in person at the reception desk. We don’t want to take over people’s jobs, but we do want to help people work more efficiently so that they have more time to spend on tasks that require a human approach.
What is the difference between this and ChatGPT?
Can’t you simply ask ChatGPT your question? The problem with ChatGPT is that it’s a statistical and generative AI tool. ChatGPT’s software can browse the data available to it in an infinitely resourceful manner, and generate a text based on this. However, ChatGPT is not capable of semantically analysing this data, nor can it assess its relevance or reliability.
To put it briefly: ChatGPT does not ‘understand’ what it generates. That’s an important difference between ChatGPT and the technology used by Nalantis, which is analytic and based on significance. Rather than just recognising individual words, our technology understands these words in their semantic context. As a result, it works on a far deeper level, and more accurately.
In an ideal world, we would merge the capabilities of our software with those of ChatGPT. At present, our application for municipal bodies only displays the found documents, and applicants still have to comb through them to find the answer they are looking for. If we were to link a generative AI tool such as ChatGPT to this, applicants would receive well-written answers that would help them immediately proceed with whatever it was they were doing when they were asking their question. Reliable data as well as smooth interaction with a chatbot: perhaps this is the ideal combination that will help us on the way to genuine openness of administration.
Would you like more information about our technology? And would you like to genuinely achieve openness of administration in your town, city or municipality? Let’s talk.
Written by Frank Aernout, CEO of Nalantis. Connect with Frank on LinkedIn.
