Google’s gunning for Duolingo with a new Google Search feature designed to help people practice — and improve — their English skills.

Rolling out over the next few days for Search on Android devices in Argentina, Colombia, India, Indonesia, Mexico and Venezuela, with more countries and languages to come in the future, the new feature will provide interactive speaking practice for language learners translating to or from English, Google said in a blog post.

“Google Search is already a valuable tool for language learners, providing translations, definitions, and other resources to improve vocabulary,” reads the the post, attributed to Google Research director Christian Plagemann and product manager Katya Cox. “Now, learners translating to or from English on their Android phones will find a new English speaking practice experience with personalized feedback.”

The new experience presents Search users with prompts and asks them to speak the answers using a provided vocabulary word. During each practice session, which last 3 to 5 minutes, Search gives personalized feedback — and the option to sign up for daily reminders to keep practicing and level up to the next difficulty level.

How personalized is it, exactly? Well, according to Google, the experience gives semantic feedback — indicating whether a response was relevant to a given question and comprehensible to a conversation partner. It also recommends areas where grammar could be improved, and, to give concrete suggestions for alternative ways to respond, provides a set of example answers at varying levels of language complexity.

During practice sessions, learners can tap on any word they don’t understand to see a translation of that word that considers its context.

“Designed to be used alongside other learning services and resources, like personal tutoring, mobile apps and classes, the new speaking practice feature on Google Search is another tool to assist learners on their journey,” Plagemann and Cox write.

These features required a fair bit of AI and machine learning engineering, Google says.

For example, the Google Translate team had to develop a model, Deep Aligner, to connect different words that create meaning to suggest translations. Other research groups at Google adapted grammar correction models for text to work on speech transcriptions, specifically for users with accented speech. Google Research teams created a separate model to power the semantic feedback component of the experience. And those same teams built another model to estimate the complexity of a sentence, phrase or individual words to “challenge learners appropriately for their ability levels.”

Google says that it recruited “linguists, teachers and ESL/EFL pedagogical experts” to craft the Search experience, yielding a mix of human-expert content — e.g. prompts, focus words and example answers — in addition to content created with AI assistance and in-house “human review.”

Other, unnamed language learning partners participated, as well, Google says — helping to surface content they’re creating with learners. And Google plans to broaden the program to additional partners in the future.

“We look forward to expanding to more countries and languages in the future, and to start offering partner practice content soon,” Plagemann and Cox continue. “With these latest updates, which will roll out over the next few days, Google Search has become even more helpful.”

One wonders what Google’s endgame might be with the rollout of the new AI-powered language learning experience for Search. Certainly, it’ll boost engagement — or could, in theory. But is it meant to lay the groundwork for a true challenger to language learning apps like Duolingo, Memrise and Babbel? The blog post’s language implies that it isn’t. Given the profits to be made in the massive field of language education, though, it’s tough to say for certain.

Source link