Presented By
ROAM Lab/Columbia Engineering

Robots have revolutionized large parts of our lives—but they sweat the small stuff. Ask a bot to handle an item like a tennis ball or TV remote, and it struggles with dexterity. Researchers at Columbia University, led by associate professor Matei Ciocarlie, are addressing that issue. They’ve created a robotic hand that, in addition to 15 independently actuating joints, has sensitive touch sensors at the end of its “fingers.” An AI learning algorithm takes in touch data and analyzes how the object is likely to move in the hand’s grip, before sending signals to the joints to move. “You bring all of these things together, and that’s how you get a really dexterous robot hand,” says Ciocarlie.

More Must-Reads From TIME

Contact us at

Hydration Made Fun
A Modern Flip Phone
A Saliva Pregnancy Test
More Than a Fast Car
A Breakthrough Bassinet