Google diversifies its translation service, adds 24 languages
This week Google added 24 languages to its translation service, including South American, African, and South Asian languages not commonly found on popular tech products. It could help keep languages alive, but certain cultural nuances may not translate.
A student colors in a fox during a Quechua Indigenous language class focusing on animal names at a public primary school in Licapa, Peru, Sept. 1, 2021. Google just added Quechua, spoken by about 10 million people, to its translation service.
Martin Mejia/AP/File
Lima, Peru
About 10 million people speak Quechua, but trying to automatically translate emails and text messages into the most widely spoken Indigenous language family in the Americas was long all but impossible.
That changed on Wednesday, when Google added Quechua and a variety of other languages to its digital translation service.
The internet giant says new artificial intelligence technology is enabling it to vastly expand Google Translate鈥檚 repertoire of the world鈥檚 languages. It added 24 of them this week, including Quechua and other Indigenous South American languages such as Guarani and Aymara. It is also adding a number of widely spoken African and South Asian languages that have been missing from popular tech products.
鈥淲e looked at languages with very large, underserved populations,鈥� Google research scientist Isaac Caswell told reporters.
The news from the California company鈥檚 annual I/O technology showcase may be celebrated in many corners of the world. But it will also likely draw criticism from those frustrated by previous tech products that failed to understand the nuances of their language or culture.
Quechua was the lingua franca of the Inca Empire, which stretched from what is now southern Colombia to central Chile. Its status began to decline following the Spanish conquest of Peru more than 400 years ago.
Adding it to the languages recognized by Google is a big victory for Quechua language activists like Luis Illaccanqui, a Peruvian who created the website Qichwa 2.0, which includes dictionaries and resources for learning the language.
鈥淚t will help put Quechua and Spanish on the same status,鈥� said Mr. Illaccanqui, who was not involved in Google鈥檚 project.
Mr. Illaccanqui, whose last name in Quechua means 鈥測ou are the lightning bolt,鈥� said the translator will also help keep the language alive with a new generation of young people and teenagers, 鈥渨ho speak Quechua and Spanish at the same time and are fascinated by social networks.鈥�
Mr. Caswell called the news a 鈥渧ery big technological step forward鈥� because until recently, it was not possible to add languages if researchers couldn鈥檛 find a big enough trove of online text聽鈥� such as digital books, newspapers, or social media posts聽鈥� for their AI systems to learn from.
U.S. tech giants don鈥檛 have a great track record of making their language technology work well outside the wealthiest markets, a problem that鈥檚 also made it harder for them to detect dangerous misinformation on their platforms. Until this week, Google Translate was offered in European languages like Frisian, Maltese, Icelandic, and Corsican聽鈥� each with fewer than 1 million speakers聽鈥� but not East African languages like Oromo and Tigrinya, which have millions of speakers.
The new languages will roll out this week. They won鈥檛 yet be understood by Google鈥檚 voice assistant, which limits them to text-to-text translations for now. Google said it is working on adding speech recognition and other capabilities, such as being able to translate a sign by pointing a camera at it.
That will be important for largely spoken languages like Quechua, especially in the health field, because many Peruvian doctors and nurses who only speak Spanish work in rural areas and 鈥渁re unable to understand patients who speak mostly Quechua,鈥� Mr. Illaccanqui said.
鈥淭he next frontier, or challenge, is to work on speech,鈥� said Arturo Oncevay, a Peruvian machine translation researcher at the University of Edinburgh who co-founded a research coalition to improve Indigenous language technology across the Americas. 鈥淭he native languages of the Americas are traditionally oral.鈥�
In its announcement, Google cautioned that the quality of translations in the newly added languages 鈥渟till lags far behind鈥� other languages it supports, such as English, Spanish, and German, and noted that the models 鈥渨ill make mistakes and exhibit their own biases.鈥� But the company only added languages if its AI systems met a certain threshold of proficiency, Mr. Caswell said.
鈥淚f there鈥檚 a significant number of cases where it鈥檚 very wrong, then we would not include it,鈥� he said. 鈥淓ven if 90% of the translations are perfect, but 10% are nonsense, that鈥檚 a little bit too much for us.鈥�
Google said its products now support 133 languages. The latest 24 are the largest single batch to be added since Google incorporated 16 new languages in 2010. What made the expansion possible is what Google is calling a 鈥渮ero-shot鈥� or 鈥渮ero-resource鈥� machine translation model聽鈥� one that learns to translate into another language without ever seeing an example of it.
Facebook and Instagram parent company Meta introduced a similar concept called the Universal Speech Translator last year.
Google鈥檚 model works by training a 鈥渟ingle gigantic neural AI model鈥� on about 100 data-rich languages, and then applying what it鈥檚 learned to hundreds of other languages it doesn鈥檛 know, Mr. Caswell said. 鈥淚magine if you鈥檙e some big polyglot and then you just start reading novels in another language, you can start to piece together what it could mean based on your knowledge of language in general,鈥� he said.
He said the new group ranges from smaller languages like Mizo, spoken in northeastern India by about 800,000 people, to more widely spoken languages like Lingala, spoken by around 45 million people across Central Africa.
It was more than 15 years ago聽鈥� in 2006聽鈥� that Microsoft got some positive attention in South America with a software feature translating familiar Microsoft menus and commands into Quechua. But that was before the current wave of AI advancements in real-time translation.
Harvard University language scholar Am茅rico Mendoza-Mori, who speaks Quechua, said getting Google鈥檚 attention brings some needed visibility to the language in places like Peru, where Quechua speakers are still lacking in many public services. The survival of many of these languages 鈥渨ill depend on their use in digital contexts,鈥� he said.
Another language scholar, Roberto Zariquiey, said he鈥檚 skeptical that Google could make an effective language revitalization tool for Quechua, Aymara, or Guarani without closer participation from community groups in the region.
鈥淟anguages are deeply linked to lives, to cultures, to ethnic groups and political organizations,鈥� said Mr. Zariquiey, a linguist at the Pontifical Catholic University of Peru. 鈥淭his should be taken into account.鈥�
Note: The new languages added are: Assamese, Aymara, Bambara, Bhojpuri, Dhivehi, Dogri, Ewe, Guarani, Ilocano, Konkani, Krio, Lingala, Luganda, Maithili, Meiteilon (Manipuri), Mizo, Oromo, Quechua, Sanskrit, Sepedi, Sorani Kurdish, Tigrinya, Tsonga, and Twi.
This story was reported by The Associated Press. Matt O鈥橞rien reported from Providence, Rhode Island.
Editor鈥檚 note: For more on how language is connected with identity, check out our聽鈥淪ay That Again?鈥� podcast.