Carnegie Mellon researchers are reimagining home automation by giving ordinary items—coffee mugs, staplers and trivets—small motorized bases and linking them to a central AI system. Using overhead cameras, computer vision and large language models, the setup predicts user intent and moves objects into position—like rolling out a trivet when someone lifts a hot baking tray. The approach sidesteps the technical and safety challenges of humanoid robots by distributing intelligence to the environment, but it raises fresh concerns about privacy, security and the wisdom of mobilizing potentially dangerous tools. Researchers say on-device processing and stronger consumer protections could address trust issues. While the technology is close to feasible, its adoption will hinge on whether households accept cameras and networked control in private spaces.
Related articles:
Domestic robot
News | Stanford Institute for Human-Centered AI
