Brain-controlled Robots

MIT developed a system to control robots with electroencephalogram signals. The robot is tasked to sort objects, while the subject wearing an EEG cap watches. The robot is looking for an error-related potential (ErrP) signal that a human brain emits when the person sees the robot about to commit an error. That's why the robot appears to make a selection and then pause to see if it's correct before dropping the object in a basket.

(YouTube link)

While the robot itself is not all that impressive, the interface between the brain and the action is almost supernatural. That an EEG can actually sort these kinds of messages is amazing. Read about the experiment in the original MIT paper.