A startup called Laplace Intelligence tests robots and self-driving cars before they face the real world. This kind of testing is called red-teaming. It means trying hard to make a system fail on purpose. If the system breaks, engineers can fix it early. Laplace is based in Ann Arbor, Michigan. Its boss, Howie Sun, says robots need this kind of safety check. The company is presenting at TechCrunch Disrupt in San Francisco.
Mr Sun studied self-driving car safety for eight years at university. In 2023, he and his adviser Henry Liu published a paper on testing this kind of AI. They started the company a year ago. The team believes that the best way to make robots safe is to have AI test them. The company is one of 20 startups in a contest at this week's TechCrunch Disrupt. Safety worries have grown as AI moves from screens into machines.
Robots and driverless cars will soon work near people. Testing them first can prevent real-world accidents and injuries.

A startup called Laplace Intelligence tests robots and self-driving cars before they face the real world. This kind of testing is called red-teaming. It means trying hard to make a system fail on purpose. If the system breaks, engineers can fix it early. Laplace is based in Ann Arbor, Michigan. Its boss, Howie Sun, says robots need this kind of safety check. The company is presenting at TechCrunch Disrupt in San Francisco.

Mr Sun studied self-driving car safety for eight years at university. In 2023, he and his adviser Henry Liu published a paper on testing this kind of AI. They started the company a year ago. The team believes that the best way to make robots safe is to have AI test them. The company is one of 20 startups in a contest at this week's TechCrunch Disrupt. Safety worries have grown as AI moves from screens into machines.

Robots and driverless cars will soon work near people. Testing them first can prevent real-world accidents and injuries.

πŸ“° Source: News Source
techcrunch.com β†—
Was this article useful?