These additional languages, including Cantonese, NKo, and Tamazight, will cater to over 614 million speakers, thereby enabling translations for approximately 8 percent of the global population.
The company stated that the newly introduced languages range from major world languages with over 100 million speakers to those spoken by small Indigenous communities, and even a few with minimal native speakers but active revitalization efforts.
Notably, a quarter of the new languages originate from Africa, marking the most extensive expansion of African languages yet, such as Fon, Kikongo, Luo, Ga, Swati, Venda, and Wolof. In 2022, Google incorporated 24 new languages using zero-shot machine translation, where a machine learning model learns to translate into another language without prior exposure to examples.
Additionally, the company shared its commitment to the "1,000 languages initiative," which aims to develop AI models to support the 1,000 most spoken languages worldwide. Among the fresh additions, Punjabi (Shahmukhi) stands out as the variation of Punjabi scripted in Perso-Arabic (Shahmukhi) and is the predominant language in Pakistan.
Noteworthy is Tok Pisin, an English-based creole serving as the lingua franca of Papua New Guinea. Google suggests English speakers attempt translating into Tok Pisin to potentially discern the intended meaning.