Or activate the microphone, and send the recording to a speech-interpreting API to see if the user has correctly recited "I'm not a pheasant plucker, I'm a pheasant plucker's son, and I'm only plucking pheasants until the pheasant plucker comes"?