Google Translate celebrates its 20th anniversary with pronunciation practice feature

Google Translate welcomed its 20th anniversary with a new feature after its launch on April 28, 2006. Google is marking this milestone with the addition of a pronunciation practice tool to the Android platform. This innovation, which has been requested by users for a long time, expands the translation experience from being limited to text only to speaking practice. The new feature aims to provide a more interactive use, especially for foreign language learners.

The new tool can be accessed through the “Practical” option that appears after the translation result is obtained. In this section, users can say the relevant expression out loud by tapping on the “Pronunciation” option. The system analyzes the speech with the support of artificial intelligence and provides instant feedback. This structure works with Gemini-based talents such as “Understand” and “Ask” announced earlier in the year to create a more holistic experience. Thus, users not only see the correct translation but also have the opportunity to learn the correct pronunciation.

Google Translate continues its development with artificial intelligence

The pronunciation practice feature is currently available on Android devices in English, Spanish, and Hindi for users in the US and India. Google states that this scope will be expanded over time. In addition, the company also shared some technical details about the history of the Translation service. When the service was first introduced in 2006, the system was based on statistical machine learning methods and large-scale language models were studied to improve translation quality. These models focused on producing more natural results by analyzing the frequency of use of words and short phrases.

In 2016, Google made a remarkable transformation by switching to neural networks. This approach resulted in smoother results that took context into account rather than word-for-word translation. This change, which showed that deep learning techniques can be applied on a large scale, formed the basis of today’s advanced translation infrastructure. In current systems, Gemini models come into play and produce faster and more accurate results, especially in real-time speech translations.

Today, Google Translate serves approximately 1 billion monthly users. The application works in integration with different services such as Google Search, Lens and Circle to Search, ensuring that a total of 1 trillion words are translated every month. More than a third of live translation sessions last five minutes or longer, indicating that users are increasingly using these tools.

The new pronunciation feature responds to a practice-oriented need within this broad ecosystem. Despite this, it is noteworthy that language support is currently limited and only available on the Android platform. Still, this approach, which provides real-time feedback, is expected to contribute to language learning processes. While Google Translate has gone beyond text translation over the years and become a more comprehensive communication tool, this added feature further expands its usage area.