New AI Device Allows Speech Without Vocal Cords
UCLA engineers have developed a wearable device that helps individuals with voice disorders regain speech. The soft, stretchy patch, about 1.2 inches wide and 0.06 inches thick, adheres to the throat and uses machine learning to translate laryngeal muscle movements into audible speech with 95% accuracy.
The device has two components: one detects muscle signals and converts them into electrical signals, while the other transforms these into voice output. Made of biocompatible materials, it is non-invasive, reusable, and lightweight.
Tested on healthy adults, the system accurately predicted sentences spoken aloud or voicelessly. The team plans to expand its vocabulary and test it on individuals with speech disorders, offering a promising alternative to existing invasive or uncomfortable solutions.