The government wants to use the capabilities of artificial intelligence to better help citizens in their use of public services.
Artificial intelligence is seen with a rather worried eye by the political powers, and in particular by the countries of the European Union, where attempts at regulation are multiplying. But states are also thinking about using technology to improve their mission to the population.
A 6 month experiment
Artificial intelligence is not reserved for private giants like Microsoft or Google. The French government proves it, with an announcement that has just been made to AFP. ” AI will inevitably transform the interactions between users and administrations », explain relatives of the Minister of Transformation and Public Service, Stanislas Guerini.
A 6-month experiment will in this context be launched from next September, with several major language models that will be trained in order to be able to respond to users of public services. Once this period has passed, a potential generalization could then be decided on the whole administration in 2024.
ChatGPT, LLaMA… and a Frenchman!
It must be said that if the AIs are impressive, they still carry their share of errors. And putting a technology that still has the flaws we know about it directly in front of citizens should not improve the credit of the state. ” Content generation models will inevitably make mistakes, we will have to train them little by little, that we perfect them, to have the most precise answer possible. “, Details the same source.
And for this interesting mission, the government will use what is currently best, since the Facebook model (LLaMA) and ChatGPT technology will be used. And following this duo of giants, a more national solution with the French company Bloom will also be used. So, is this an idea for the future, or an experiment that is too technical and too risky to finally succeed?
Source : Provence
17