Chinese researchers develop innovative textile that helps AI recognize voice commands

NANJING -- AI can empower traditional industries while traditional technologies can, in turn, advance AI, as shown in a new study where an innovative textile helps AI recognize voice commands by amplifying the electrostatic charges generated during speech.

Published this week in Science Advances, the breakthrough promises to turn garments into intuitive, always-accessible AI assistants.

Led by researchers from Soochow University, the team designed a triboelectric acoustic textile, dubbed A-Textile. Unlike conventional rigid and bulky acoustic devices, the fabric is soft, flexible and washable.

It can harness the natural electrostatic charges generated on clothing when a person speaks. The researchers enhanced this effect by creating a multi-layered structure featuring a composite coating of 3D tin-sulfide nanoflowers embedded in silicone rubber, along with graphite-like carbonized textile.

The washable and lightweight AI-textile achieved a voice recognition accuracy of up to 97.5 percent, enabling remote control of household appliances, according to the study.

The researchers demonstrated its practical application by using it to wirelessly control smart home appliances, such as turning on and off an air conditioner and a lamp through simple voice commands.

It also successfully accessed cloud-based services, using voice commands to operate smartphone applications like Google Maps for navigation, and engage with ChatGPT by asking for cocktail recipes and a travel itinerary.