A team of scientists at New Scientist has discovered the first fully artificial voice.
The team says it’s a “huge breakthrough”, and they believe it will help people with complex medical conditions.
The breakthrough comes from a project called the HumanVoice.
The HumanVoice is a voice-tracking app that helps doctors and patients communicate better and is used by more than 10,000 medical facilities around the world.
It’s also used by millions of people around the globe to monitor their health.
A team led by Dr Andrew Alder at University College London and colleagues has developed a voice recognition app that can learn from other voices, to recognise a patient’s speech patterns and use that to make decisions.
This has applications beyond speech recognition, for example to help patients and other healthcare professionals recognise a loved one.
Alder says that, while there are other approaches to artificial voice recognition, “The HumanVoice has been an extraordinary success in the field of voice recognition.”
It’s a significant advance, and Alder believes that it’s “a big step forward”.
The team first tried out the app in a human voice in 2014.
In 2015, they used a “human” voice to help doctors diagnose a patient with Parkinson’s disease, and in 2018, they applied it to analyse a patient and to make an emergency decision about whether or not to keep them in hospital.
The human voice has a lot more processing power than what is usually possible with voice recognition apps, but the team says the new app can learn to recognise people’s speech from a wider range of voices than they have previously been able to.
So, for instance, it can recognise a woman’s voice and recognise her as a person.
The technology is still relatively new, and the team’s next step is to develop an app for other types of speech, such as spoken word, which they plan to do in the next few years.
The project is funded by the UK’s National Health Service.