The famous translation tool, which already offers 133 languages, will add 110, including regional languages.
An extraordinary polyglot. Google translate added 110 languages to its repertoire this Thursday, on the occasion of its largest “great addition ever made”. including Breton and Occitan. “These new languages represent more than 614 million speakers, paving the way for translations for approximately 8% of the world’s population”welcomes the American Tech giant.
Among the added options are regional languages, including Breton and Occitan. Cantonese, Tamazight, Romani, Wolof, Panjabi and Q’eqchi’ are also integrated. “Some are major languages in the world, with more than 100 million speakers. Others are spoken by small indigenous communities, and a few have almost no native speakers but are making efforts at revitalization.”the company said in a blog post. Languages such as Punjabi, the most widely spoken language in Pakistan, and Afar, a language spoken in Djibouti, Eritrea and Ethiopia, are available.
“The project lasted four years” tells our colleagues at Ouest France Isaac Caswell, software engineer for Google Translate. “This update is the largest in Google Translate’s history and extends to many more non-national languages and disenfranchised languages. It also includes more Creole languages than ever before, including French-derived Seychellois Creole and Mauritian Creole.
New languages have been chosen “based on a combination of requests received, the number of speakers and the amount of data available for the training models” says Isaac Caswell. Catalyzed by the PaLM2 language model, the basis of its Gemini generative AI, Google used artificial intelligence to train its translation tool. “We consulted linguists, experts, translators and native speakers. They assessed the quality of the model and provided translations”adds the expert.
Breton demands and initiatives
Last February, a Breton collective campaigned for the integration of the Breton dialect into Google Translate. A step “essential for Breton and Brittany to count in the world” explained David Lesvenan, president of the Bretagne digital endowment fund. The public office for the Breton language had also set up a historical dictionary and a toponymic database. These tools aim to establish the Breton language in the digital sphere with more than 60,000 words translated.
Read alsoAdvertising: Google targeted by proceedings worth 13.6 billion pounds in the United Kingdom
After adding 24 new languages in 2022 with Zero-shot technology, which allows a learning model to translate a language without relying on examples and therefore better personalize the rendering, Google launched a plan, called the thousand languages initiative. Objective : “create AI models that will support the 1000 most spoken languages in the world”.
