Google Translate Adds 110 Languages, Expands Reach to 614M More People Worldwide

Google Translate has significantly expanded its capabilities by leveraging Google’s PaLM 2 large language model, allowing it to accurately translate a much wider array of languages, dialects, and creoles.

Key details of the expansion: Google Translate is adding 110 new languages, the largest single expansion in the platform’s history, enabling it to serve an additional 614 million people, or about 8% of the global population:

• The update includes some widely spoken languages like Cantonese and Punjabi (Shahmukhi), as well as lesser-known tongues like Manx, Sicilian Italian, and Tok Pisin.
• African languages saw their biggest expansion, with the addition of Fon, Kikongo, Luo, Ga, Swati, Venda, and Wolof.
• For languages with regional variations, Google Translate aims to provide translations that can be widely understood, such as offering a Romani language option that includes three different dialects.

How PaLM 2 enables the expansion: The integration of Google’s advanced PaLM 2 large language model is key to Google Translate’s ability to handle this broader set of languages with improved accuracy and cultural relevance:

• PaLM 2 enhances Google Translate’s efficiency in learning and shifting between languages, particularly closely related ones like Hindi and its dialects, or various French creoles.
• Originally developed to facilitate human-AI communication, PaLM 2’s adaptability allows it to manage nuances and variations within languages more effectively.

Accessibility and implications: The expanded language capabilities are now available to users via the Google Translate website or mobile apps on Android and iOS, significantly increasing access to translation services for millions of people worldwide.

Broadening communication horizons: The addition of over 100 new languages to Google Translate marks a major milestone in making translation technology more inclusive and accessible to a wider global population. By leveraging advanced AI language models like PaLM 2, Google is opening up new opportunities for cross-cultural communication and understanding, particularly for speakers of lesser-known languages and dialects that have previously been underserved by translation platforms. However, it remains to be seen how well the AI model handles the complexities and nuances of these newly added languages in practice, and whether the translations will truly resonate with native speakers.