Speechmatics achieves AI breakthrough, beating tech giants in tackling bias and inclusion to understand all voices

The Cambridge-based company Speechmatics has launched its 'Autonomous Speech Recognition' software, which uses self-supervised learning techniques and deep learning to improve the accuracy of speech recognition technology. The technology reduces speech recognition errors for African American voices by 45% compared to Amazon, Apple, Google, and Microsoft, achieving an overall accuracy of 82.8%. This breakthrough enables a more comprehensive representation of all voices and dramatically reduces AI bias and errors in speech recognition. The software also outperforms competitors on children's voices, recording 91.8% accuracy. Speechmatics' technology is trained on huge amounts of unlabelled data from the internet, such as social media content and podcasts, to deliver a far more comprehensive representation of all voices.