Breakthrough: AI stickers aid those with voice disorders to speak

A groundbreaking adhesive patch, integrating artificial intelligence (AI), has been developed to assist individuals with voice disorders in regaining their ability to speak. This innovative patch deciphers the movements of throat muscles and transforms them into audible speech.

Crafted by researchers in the United States and detailed in a publication in Nature Communications, this compact patch is energized solely by the movements of the user’s muscles. Once applied to the skin, it captures the larynx muscle motions, converting these into electrical signals. These signals are then processed by a machine learning algorithm, an AI form, to generate speech.

The device is capable of vocalizing specific pre-set sentences through a speaker, based on the intended speech of the user. Lead researcher Ziyuan Che from the University of California, Los Angeles, highlighted that the device operates without requiring the vocal cords, making it a potential solution for those with vocal cord damage.

Voice disorders affect nearly a third of individuals at some point in their lives, according to the study. Traditional remedies for severe voice issues, such as using handheld devices like electrolarynxes, often prove to be cumbersome, uncomfortable, or invasive.

The waterproof patch, roughly the size of a large coin and weighing a mere seven grams, resembles the texture of a rubber glove. In trials involving eight participants without voice disorders, the device demonstrated its effectiveness across various activities like standing, walking, and running. It successfully interpreted the spoken and whispered sentences with about 95% accuracy.

The patch occasionally struggled with words or names that caused similar throat muscle movements, such as “make” and “Mark”. However, Che explained that these words typically appear within broader sentences, which helps the device to understand the context and distinguish between them more effectively.