Paris Martineau, the Outline:
Why do voice assistants need to talk so much? If you’ve ever used one of Amazon’s ridiculous, yet rather addictive (I have two) Echo products, you know what I’m talking about: Whether you’re setting a timer, or asking her to play a podcast, Alexa just won’t shut the fuck up. Even when you give it a relatively simple command (like, “Alexa, set an alarm for 6 a.m.,” or “Alexa, set timer for five minutes” it always responds with either a partial or total repetition of your phrase (“Okay, alarm set for 6 a.m. tomorrow,” or “Timer set for five minutes”), which can be more than a little annoying when it’s two in the morning and you don’t exactly want a booming robot voice waking your roommates up a wall over.
Siri does this too, but it’s smart when used on an iPhone: voice feedback is much less verbose if you activate Siri by using a hardware button instead of saying “Hey, Siri”. Of course, it’s only able to be smarter because it has a screen.
Audio-based feedback is helpful for confirming requests on a screen-less voice-driven interaction, but I often wish these replies could be faster and less wordy. As Martineau says, it is a bit annoying when the virtual assistants get the request right, but I think it’s even more irritating when something is interpreted incorrectly. I’m sure there are some users who love the attempt at personality, but I’m sure I’m not the only one who would love a sliding scale where I could reduce its veneer of humanity. More generally, I’m not convinced that attempts at anthropomorphizing technology actually makes it any more useful or trustworthy.