AI development is progressing rapidly not only in the gaming sector, but also in matters of health. There are already many new achievements. This includes the development of a simple sticker that will soon help people who have lost their voice.
Those: Canva
The idea for such a sticker came from a team of developers at the University of California in Los Angeles. However, there is currently only a prototype. The small, battery-free sticker is intended to help people with various vocal fold diseases speak thanks to AI.
The current prototype looks like a sticker, is about the size of a plaster and weighs about seven grams. The small device was tested as part of a study. A picture of the prototype was also published on X.
How does the sticker work?
The device is supposed to detect the movements of the larynx, so when a patient speaks a sentence, the prototype converts the movement into electrical signals. The signals in turn are then sent to an external module. There they are analyzed and converted into language with the help of artificial intelligence. There is no loudspeaker; according to the researchers, the sound should also be reproduced via the sticker. However, there is no further information on this yet.
An article on ScienceAlert explains the experiment setup. A total of eight people were asked to test the sticker in various movement situations. They should say phrases like “I love you,” “Merry Christmas,” or “I don’t trust you” while walking or sitting. Surprisingly, the small tool did not require an external power supply. The researchers commented as follows: “Here we present a self-powered, wearable sensor activation system based on soft magnetoelasticity that enables assisted speech without relying on the vocal folds.“
AI stickers in research phase
Since the study initially only involves eight test subjects, the test does not yet offer a final result. According to the study, the device should have a voice recognition accuracy of 95 percent. Only when it came to words that sounded very similar did the sticker have difficulty telling them apart.
But there are still limitations: The device is currently only able to play back pre-recorded sentences. Nevertheless, the study’s lead author, Ziyuan Che from the University of California, Los Angeles, expressed optimism. He emphasized that the device works without detecting a person’s vocal cord vibrations, which could help people with damaged vocal cords regain their speech. Che believes that thanks to advanced algorithms, the prototype could one day skip the pre-recording step. However, it is important to emphasize that this is still a prototype and will likely require several years of development before it is ready for use in patients.