You Can Now Improve Your Pronunciation With Google AI

A new machine learning-enabled new feature of Google will help users pronoun unfamiliar words.

Here is how the feature works. When users search for a word, they can try speaking the word into a microphone and Google’s AI will analyze their pronunciation. They will then receive feedback on how each syllable matches what Google expects the word is pronounced.

Let’s take the word “asterisk” as an example. Google will use speech recognition technology to analyze users’ pronunciation, and then it will tell users that the last syllable is pronounced “rict,” not “uhsk.” With this feedback, users can improve the next time they say the word.

The feature has already been available but for now, it is still at the experimental stage and users can only access it on mobile devices. For now, the function can recognize American English only but the company will launch the same feature for Spanish soon.

Google is also trying to improve its ability to translate words with a visual aid. For example, if you want to know what “naranja” means in English, you will see the word “orange” along with the picture of the fruit. Another example is “seal,” a word with many different meanings. Users will see pictures of the mechanical seals, the marine mammal, and embossed pieces of wax.

According to Google, for now, the visual aid will work in English only and for nouns, the type of words that is the easiest to visualize. However, the coverage will be expanded shortly.

The two new features help to turn Google from a simple search engine to a language coach. With the aid from machine learning and the wealth of data from users, we may expect the features to get better over time.