AWS Unveils Generative AI-Powered Transcription Platform

AWS has added new languages to its Amazon Transcribe product, offering AI-based generative transcription for 100 languages and a host of new AI capabilities for customers. Amazon Transcribe, announced during the AWS re:Invent event, can now recognize more spoken languages and trigger call transcriptions. AWS customers use Transcribe to add speech-to-text capabilities to their AWS cloud applications.

In a blog post, the company said Transcribe was trained on “millions of hours of unlabeled audio data from more than 100 languages” and uses self-supervised algorithms to learn human speech patterns in different languages and accents. AWS said it ensured some languages were not overrepresented in the training data to ensure that less-used languages could be as accurate as more frequently spoken ones.

Amazon Transcribe supported 79 languages as of the end of 2022. According to AWS, Amazon Transcribe is between 20 and 50 per cent accurate in many languages. It also offers automatic scoring, custom vocabulary, automatic language identification, and custom vocabulary filters. It can recognize speech in audio and video formats and noisy environments.

With better language recognition, AWS said advances with Amazon Transcribe also translate to greater accuracy with its Call Analytics platform, which its contact centre customers frequently use. Amazon Transcribe Call Analytics, now powered by generative AI models, summarizes interactions between agents and customers. AWS said this reduces post-call work to create reports, and managers can quickly read information without needing to review the entire transcript.

AWS also announced additional capabilities for its Amazon Personality product, allowing customers to offer products or show recommendations to customers. For example, streaming services can suggest new shows based on previous activity. AWS added Content Generation, which will write email titles or subject lines to thematically connect recommendation lists.