Thoughts on Voice-Input

I have only seen people use voice-input on their devices rarely.

Recently I have begun using Siri on my iPhone quite regularly. In the past I didn’t do it, as at best voice-recognition of commands feels rather gimmicky, and only works 90% of the time.

I started using it as a joke, saying “Siri, Shutdown” when leaving work before manually powering down my work-mac as there is no actual shutdown-voice-command (for safety-reasons, one assumes).

But there are some times when I DO use voice-commands, and which are now a preferred way of mine of doing certain things. These are:

  • Getting the time while on a bike (using my headphones to interact with the phone)
  • Setting a timer in my elevator after starting the communal washing-maschine
  • Taking a note while on a bike (using my headphones to interact with the phone)
  • Getting the dictionary-definition of a word

That’s it, mostly. Note that these are all actions which

  • Would require a non-trivial amount of button-presses
  • I do when I do not or cannot comfortably access my phone, but have headphones with a remote on
  • I do while being alone

Even when in company I would rather not use them, as I tend to speak less loudly and clear, the voice-recognition will invariably fail, and for not wanting to annoy people in my surroundings (akin to yelling “I am on the bus!” into a phone while you are on the bus).

One case I remember is me playing the game Endwar at game-design-school, whose gimmick was using voice-command to control your army. Everyone else hated it, as it meant several people speaking loudly into their headsets, annoying everyone else. People hated it even when I spoke the word “three” 14 times in a row as the system failed to recognize it (Siri cannot pick up my “three” reliably either).

It concludes that people are afraid of using voice-inputs as they are

  • Embarassing
  • Irritating to others
  • Only accurate when spoken loudly
  • Only accurate 90% of the time


I like using voice-commands when they save taking out my phone, would have to be done with a lot of actions, on my bike, and when I am alone.

I feel concentrating on these actions and teaching people on the possibilities would make them easier for people to understand and use, as they come with direkt, tangible benefits.

Once a baseline of feasible voice-commands is established these can move beyond the current stage and develop into things that cannot be anticipated now.