Amazon Transcribe Presents 21 New Languages With Generative AI Support

Amazon Transcribe presents a new language with AI support (photo: dock. Amazon)

JAKARTA Amazon Web Services (AWS) has added a new language with the support of Artificial Intelligence (AI) generative to Amazon Transcribe products, voice-to-text transcript services.

Citing from the Engadget report, Amazon Transcribe has added dozens of new languages so that the platform can recognize more spoken languages. So far, Amazon Transcribe has had 100 languages.

The presence of the new, unencangled language is supported by the use of algorithms that are directly supervised by AWS. This algorithm will study speech patterns from various languages and accents.

By continuing to train Transcribe using millions of hours of unlabeled audio data, AWS hopes that this transcript platform can provide the best service and can remain accurate regardless of the accents used by the speaker.

In 2022, Amazon Transcribe only supports 79 languages and its accuracy is still so small. The company recognizes that the accuracy of their transcript services is only 20 to 50 percent in various languages.

Currently, Amazon Transcribe is capable of including automatic reading marks, special vocabulary, automatic language identification, to filters from special vocabulary. One of the best capabilities of Transcribe is speech recognition in noisy environments.

In addition to increasing the accuracy of the language, Amazon Transribe also includes the ability of a generative AI in the Transcribe Call Anaylitics feature. With this capability, users can summarize the interaction between agents and customers more easily.