The government's goal is to make the administration more modern and efficient. Artificial intelligence should primarily take over repetitive tasks, analyse documents, simplify research, or help with translations. At the same time, Minister for Digitalisation Stéphanie Obertin said that technology should not replace humans.

She said that the government sees AI as an important part of modernising the state and is examining how it could be used within the public administration. However, the aim is not to replace staff or civil servants, she noted.

"This is about helping them make their work easier", Obertin said, particularly by simplifying administrative procedures and reducing repetitive tasks.

Now that the test phase has concluded, new AI capabilities have been made available to government employees. The artificial intelligence comes from the State's own data centres and is accessible via a chatbot, as well as other means.

In addition to active use by employees, the technology is now also to be gradually integrated into administrative processes, according to Patrick Houtsch, director of the Government IT Centre (CTIE).

He said AI could be used to handle the initial processing of documents submitted through MyGuichet or by email, including categorising them and checking whether they are complete.

"These are stages that an AI can carry out", he said, adding that the technology would therefore need to be integrated into administrative processes.

AI is not only meant to speed up administrative work, but also simplify the service for citizens. At the same time, however, the question arises of how sensitive data will be protected. For the CTIE, this is one of the biggest priorities.

Houtsch stressed that protecting government data was a central concern and explained that this was why Luxembourg had opted for a sovereign AI system operated in its own data centres.

"The data does not leave the state", he said. "We know what happens to it and where it is stored."

In addition, an AI charter will prescribe how to use the technology responsibly. For example, decisions that directly affect citizens should not be made by AI alone. An employee must monitor this to ensure transparency and reliability.

The goal now is to gradually make AI accessible to citizens, in order to simplify exchanges with the administration.