This AI Can See Through Walls

Pete Cooper describes research using Wi-Fi signal reflections and AI to estimate a person’s pose through a wall. He says a Carnegie Mellon and Facebook AI research team used low-powered signals and a neural network to identify body points such as the head, arms and legs.

The video considers both the potential safety applications and the serious privacy implications of sensing people who cannot be seen directly. It frames the technology as a research development with ethical consequences, rather than a casual consumer feature. The speakers raise the question of how such capabilities should be controlled.