Robots Gain Sense of Touch in Breakthroughs at Columbia, Bristol, and Stanford

Columbia University engineers have developed a robot hand with pressure sensors that can feel fragile objects like eggs and grapes, adjusting grip in real time. Similar advances at Bristol and Stanford bring tactile sensing closer to human levels, promising safer automation in manufacturing, healthcare, and logistics across Europe.

Bottom line — Columbia's tactile system achieved human-like precision handling delicate objects, a milestone for AI in physical tasks.

Go deeper (7)

• Columbia's system uses pressure sensors that adjust grip in real time, enabling the robot to reorient tools and serve food with a spatula, per Columbia Magazine.
• At the University of Bristol, engineers built a fingertip with tactile sensitivity approaching human skin, detecting ridges smaller than a grain of sand, said project lead Nathan Lepora.
• Stanford researchers created electronic skin with 1,000 sensors per square centimetre, though the human hand has about 17,000 receptors per square centimetre, according to a technical analysis.
• The technology still faces high costs, latency issues, and no global standards for haptic hardware and protocols, notes The Data Scientist.
• Tactile intelligence is key to moving AI from screen-based recognition to physical manipulation in logistics and manufacturing, argues invisibletech.ai.
• Researchers use digital twins and NVIDIA-style simulators to train haptic models, generating millions of safe trials before real-world deployment, per invisibletech.ai.
• Applications range from robotic surgery and prosthetics to telehaptic remote control and care robots that lift elderly patients, according to multiple sources.

Read the reporting