That clunky old Android phone sitting in a drawer may be more useful than you think. A team from Carnegie Mellon University has created a program called Zensors that uses connected smartphone or surveillance cameras to track your environment, figure out what's going on and give you valuable alerts and statistics. The team showed how a user can point a smartphone outside a window, circle an area of interest, and pose a natural-language question like "how many cars are in the parking lot?" Zensor then proceeds to track cars as they enter and leave, giving a business data about its customers.
Such tasks are way beyond the means of regular smart home cameras and sensors, but the tech isn't quite as magical as it sounds. The researchers are relying on crowd-sourced workers to for complex tasks like counting cars, while easier tasks are calculated by algorithms. However, they hope that computers will gradually take on the more difficult jobs thanks to machine learning systems, eliminating the human factor.
Asking a machine to make complex observations sounds a bit far-fetched, given that IBM's supercomputer-based Watson has only just learned to make a Bruschetta. On the other hand, machine vision is getting pretty darn good, and artificial intelligence has advanced enough to scare Elon Musk and Stephen Hawking. Right now Zensor is still in beta testing, but if it ever becomes a thing, it could open up a new way of monitoring your life without sticking chips into everything.