Dual‐Functional Liquid Metal Fibers for Machine‐Learning‐Assisted Gesture Recognition and Adaptive Thermal Management

ABSTRACT Globally, over 100 million people with language barriers rely on sign language, but existing assistive technologies, constrained by environmental conditions or insufficient performance, fail to efficiently connect sign language with digital systems. This causes obstacles in education, healthcare, and employment. To address this, we developed a liquid metal‐based triaxial elastic conductive fiber (LM‐TECF) via continuous wet‐spinning. The fiber, composed of a eutectic gallium‐indium alloy dual conductive core and a thermoplastic polyurethane sheath, exhibits high conductivity (7.08 × 10 5 S/m), good mechanical softness (Young's modulus of about 0.2 ± 0.1 MPa), and excellent sensitivity (gauge factor of 0.539). In addition, the relative resistance change of the fiber under 300% tensile strain is only 86%, and there is no sharp increase in resistance or failure of the conductive path, which fully proves that it still has stable conductive properties under large deformation. Leveraging LM‐TECF, we developed a smart glove with 10 sensing channels, enabling real‐time finger kinematics monitoring (<100 ms response) and 98.9% sign‐language letter recognition accuracy with machine learning. Additionally, its electrothermal function (20°C–85°C) overcomes liquid‐metal phase‐change limitations and enhances comfort in low‐temperature settings. This dual‐functional platform provides an efficient communication tool for the hearing‐impaired and supports their social integration.