Microsoft announced its plan to produce digital data for AI models in more than 12 languages, as part of its efforts to promote linguistic diversity in AI technologies.
The initiative includes digitizing non-English books and preparing hundreds of hours of audio recordings in several languages.
The tech giant plans to open new branches of its research centers in Strasbourg, eastern France, starting next September, with the aim of expanding data availability in at least 10 of the 24 official languages of the European Union.
Microsoft president, Brad Smith explained that most current AI databases are largely trained in English, making models less effective when used in languages where data is insufficient. He added that this could push users to switch to English instead of their native language.
