1. 13

  2. 10

    Radio inducing a microphone input in the wire is clever!

    The idea I’ve had for doing something like this, but haven’t had time to test out, is to look for a false-positive signal for the voice-recognition software: a sound that you could broadcast as normal sound through loudspeakers, which the algorithms would hear as “Ok, Google”, but which humans would perceive as noise or some other kind of non-speech sound.

    1. 3

      It’s a shame the original article at IEEE Website is not open access - would have preferred to read the original research.