Civil servant Thomas and his AI journey

Thomas’ career has taken him from the analogue world to the digital age. He witnessed the move from paper archives to digital databases; the evolution from dial-up internet to continuous connectivity; and the transition from simple mobile phones to today’s multifunctional smartphones. When he thinks back on these technological leaps, he reflects on how they transformed his work.

The digital transformation not only caused a shift in how work was done, but it also opened a world of new possibilities. Although the changes often frightened him at first – or at least made him unsure – he learnt that adapting pays off. He does not mourn for what was, but embraces the innovations. Thomas sees digitalisation not only as a means to gain efficiency, but above all as a way to serve his city’s residents even better.

Like millions of others, Thomas began experimenting with ChatGPT last year. As with previous waves of digitisation, he felt curious, sceptical and overwhelmed. Would large language models and generative AI once again change his job?

Step by step, like a child clinging to furniture while taking its first steps, Thomas started with simple tasks. Each interaction with AI was a step forward: a small but significant move towards competence. With each question and answer, he grew confident that those uneven steps would not lead to a fall, and he discovered its potential.

On time-consuming and boring jobs, Thomas managed to save time; for communication assignments it proved to be a convenient tool to tailor messages to specific media and target audiences; in the absence of colleagues, ChatGPT was a valuable sounding board for inspiration or critical feedback. He also discovered that ChatGPT was great at coding and analysing, and an excellent watchdog for checking and ensuring quality criteria.

It also allowed him to spend more time on the things he excels in and draws energy from. He could focus on personal interaction with residents and organisations, while also providing more customised services.

But Thomas also felt like he was paving his own way over the past few months. Exploring and learning was a quest. The output was not always reliable and had to be checked. ChatGPT was certainly not the solution to every problem either. It was important to keep an eye out for suitable use cases, where applying the technology provided true support and not just another way of doing what Thomas could do himself. Let’s not forget the search for footing in frameworks and reference points. What data can I input into ChatGPT? Can I use output from ChatGPT just like that?

With the advent of the European AI Act, the Flemish AI Competence Center and the Flemish AI-strategy, Thomas was glad to see more guidelines appear. He is pleased that Flanders wants to play a pioneering role in the adoption of AI within the government.

As such, Thomas looks back satisfied and proud. The past year was a story of personal growth and adaptation, but also proof of how employees at local governments can serve citizens even better provided they acquire the right skills. ‘The possibilities for the future are not bad at all’, Thomas thinks. ‘More hands at the bedside in healthcare, more people in front of the classroom in education, and more time for residents.’

Translated into English by Heather Young