Google Voice-To-Text Adds 21 Languages
Google, get this down for me please: pen emoji, computer emoji, speaking emoji, fire emoji.
That’s exactly what you can tell the company’s speech-to-text feature now, as they have added 21 new languages to to their list of recognized inputs according to a blog post. The new additions include Armenian, Swahili, Lao, Nepali and of course, emoji. The total amount of languages recognized now sits at 119, but emoji is only recognized in English and in the US.
The search engine giant claims its speech-to-text feature is up to three times quicker than writing. If you want to access the feature, you need to download Gboard on your phone or use voice search in a browser. Part of this update revolves around honouring language, as they added capabilities to understand Georgian, a language that dates back over 1000 years.
Google worked with native language speakers to collect samples, then used those to train their machine-learning models. This means the system will learn new patterns and adapt over time as it becomes exposed to new speakers. The new languages will be added to the whole suite of Google offerings soon, including Translate.
These additions are all in an attempt to make the internet more inclusive, so that any sort of user can access every feature available to them. Technology companies are beginning to see developing countries like India as potential major consumers, so adding local language support is a logical step for Google to progress with.
