A British woman who lost her ability to speak due to a debilitating disease has been given her original voice back, thanks to artificial intelligence (AI) and a tiny, old recording.
Sarah Ezekiel, an artist from London, was diagnosed with Motor Neurone Disease (MND) 25 years ago. The disease gradually weakens the muscles needed for speech. For decades, she has communicated using a computer-generated voice that didn’t sound like her own. Her two children had never heard their mother’s real voice.
Recently, a company called Smartbox offered to create a synthetic version of her original voice. Usually, this requires a long, high-quality recording of the person speaking. However, Sarah only had a very short, eight-second clip from an old home video with background noise. Experts thought it might not be enough.
Using advanced AI technology from a company called Eleven Labs, the team was able to extract Sarah’s voice from that brief clip. The AI then learned to replicate her unique sound, including her London accent and even a slight stutter she used to have.
When Sarah heard the AI-generated voice for the first time, she was overwhelmed. She said she “almost cried.” The voice was a perfect match, allowing her to speak in her own voice again for the first time in a quarter of a century. For her friends and family, it was as if her real voice had returned.
This breakthrough is significant for many people with conditions like MND. It shows that even a tiny snippet of audio can be used to preserve a person’s unique identity and help them communicate in a voice that truly feels like their own.
Also read; http://artificial intelligence
