Artificial intelligence, the Italian ChatGpt debuts

There is a declared and promising Italian national champion in the artificial intelligence market. And it is indeed called Italy. The Italian company iGenius has presented it by making this model available to everyone, in collaboration with Cineca (the largest Italian computing pole, an inter-university consortium).

Italy is a 'large language model', like Gpt, which is the basis of the well-known ChatGpt. The difference is that it is designed for the needs of our country. It is explicitly aimed at our companies and PA, with the goal of helping them grow through artificial intelligence.

Italy was formed with almost only Italian data, unlike the other models, which were based on English. This has the advantage of a better understanding of our language, its nuances, and our historical and cultural context. There is also a gain in efficiency, because the other models, when they have to handle languages other than English, do continuous translation work. By collaborating with Editoriale Nazionale, a company of the Monrif group, the company was able to use their historical archive of press articles as a supplementary source to improve the model. It has nine billion parameters, so it is quite a large model: a promising factor for its quality.

Italianness is also in the spirit of the product, for its goals of serving the country system; hence the decision to allow free downloading and customisation (it is open source).

In short, the aim is to allow the country - as Cineca and iGenius stated during the presentation - to be a player in the artificial intelligence revolution, thus deriving as much economic value as possible from it; if we were to use only foreign products, we would be reduced to the rank of mere consumers of something that could soon spread to every market and sector of society (from factories to healthcare).