New Delhi, India

Researchers have created a new adhesive patch that uses artificial intelligence to help people with voice disorders speak again.

Advertisment

The AI will 'read' their throat muscles' movements and convert them into speech. 

The small adhesive patch has been powered only by the muscle movements, US-based researchers said in a study published in Nature Communications.

The patch, attached to the skin, changes the movements of the larynx muscles into electrical signals and translates it into a speech created by a machine learning algorithm, a form of AI.

Advertisment

After this, the speaker plays the sentences which a person intends to say aloud using the prototype which can currently relay only certain pre-recorded phrases.

Also Read: Workplace AI, robots, and trackers can be detrimental to quality of life, study says

Speaking to AFP, lead study author Ziyuan Che of the University of California, Los Angeles said the patch works without any need to use the vocal cords of the person.

Advertisment

This means it can help restore the speech of those with damaged vocal cords. 

As per the study, almost a third of people suffer from at least one voice disorder in their life.

The researchers said the existing treatments for serious voice issues such as handheld devices like an electrolarynx can be "inconvenient, uncomfortable, or invasive." 

Waterproof patch with 95 per cent accuracy 

The waterproof patch is almost the size of a large coin and its weight is only seven grams. Its texture is similar to that of a rubber glove.

For testing the device, the researchers used eight volunteers, who did not have voice disorders, to both speak and whisper five sentences.

It was seen that the patch worked while standing, walking and running.

Watch: Google restricts Gemini AI from answering poll-related queries

As per the study, the device was able to predict accurately what the volunteers were saying nearly 95 per cent of the time.

"Some words that move throat muscles in a similar manner — such as the word "make" and the name "Mark" – could be difficult for the patch to tell apart," said Che.

"But those two words usually appear in a long sentence like 'I am going to make dinner,' or 'How you doing Mark?'," Che added.

(With inputs from agencies)