How BBC Sounds is using GenAI to boost accessibility in audio

A new BBC Sounds pilot, running on the OpenAI Whisper transcribing model, explores how generative AI might automate subtitles for its 27,000 hours of audio content it produces a month