Particle.news
Download on the App Store

AI 'Co‑Pilot' Bionic Hand From University of Utah Raises Success Rates in Lab Trials

The shared-control system seeks to ease the cognitive load blamed for high prosthesis abandonment.

Overview

  • University of Utah researchers reported the results in Nature Communications on Dec. 9, detailing an AI-guided prosthetic that augments user intent rather than taking over.
  • Silicone-wrapped fingertip pressure and proximity sensors feed an AI controller that selects grasp postures and adjusts force with individual-finger control.
  • In controlled tests involving four amputee volunteers, task success rose from roughly 10–20% without the AI to about 80–90% with it, while users reported less mental effort.
  • The design preserves user agency so wearers can tighten, loosen, or release a grip at will, though researchers note it is not as dexterous as a natural hand.
  • The team plans home and real-world trials, exploration of neural interfaces to improve control and feedback, and partnerships for larger clinical studies and commercialization.