Google has announced a significant update to Google Translate, incorporating 110 new languages into its system. This marks the most extensive expansion in the service’s history, previously supporting 133 languages. The integration of these languages leverages Google’s cutting-edge PaLM 2 AI language model, renowned for its capability in mastering languages that share similarities.
This update particularly benefits languages closely related to Hindi, such as Awadhi and Marwadi, as well as various French creoles including Seychellois Creole and Mauritian Creole, as highlighted in a blog post by Google’s Isaac Caswell. The addition of Cantonese, a language frequently requested by users due to its complex relationship with Mandarin in written form, underscores the technological strides made in language processing.
Caswell notes that the challenge of training models for Cantonese was significant due to the overlap in writing with Mandarin, which complicates data collection and model training. Moreover, approximately one-fourth of the new languages are from Africa, broadening the linguistic diversity and accessibility of Google Translate.
The impact of this update is considerable, with most of the newly supported languages spoken by communities numbering at least one million. Some languages in the update are used by hundreds of millions, suggesting a dramatic increase in the utility of Google Translate for global communication.
This expansion not only enhances the functionality of Google Translate but also significantly broadens the scope of linguistic accessibility, fostering better global communication and understanding.