So Google I/O kicked off, and as expected, Google Assistant was the star of the show. Many of the exciting developments coming to Assistant lean into making our interactions with Assistant more 'human'. One of the weirder ones is that Google would appreciate if you mind your manners while speaking to it, thank you very much.
The reason for this seems wholesome enough. It's intended for family situations with kids present. Assistant's Pretty Please feature makes it reward users for issuing voice commands while also saying “Please” by praising them with responses like “thanks for saying please,” or “what a nice way to ask me.”
Need to catch up on the latest from Google? Check these highlights from I/O that will change your life:
At the event, Google's Scott Huffman, VP of engineering for Google Assistant, stressed that children would be using Assistant to ask questions, play games, and listen to stories — (families have played over 130,000 hours of stories to their kids via Assistant in the last two months alone).
Apparently, this led to concerns among the Google team that their kids would learn bad manners from bossing Assistant around all day, and so Pretty Please was born. It's an optional feature, so the less refined among us can rest easy. Assistant isn't able to reprimand us for not being polite, either.
Nonetheless, behind all the good behavior, there's something that rubs me the wrong way about being deferential towards Assistant. As useful as Google Assistant is, treating it like it has feelings that could be hurt, or hearing that it feels grateful when I know it doesn't, gives me that uncanny valley feeling.
Does Assistant deserve respect?
Google wants our interactions with Assistant to be more natural, to feel more 'human'. To a degree, this also entails accepting some kind of status of 'personhood' for Google's representative in our social interactions. Politeness, deference, manners, being considerate and so on.
The dark side, to this, is that Assistant isn't a person. It's a machine, a really helpful machine that we pay for largely by surrendering our our most intimate secrets to it in exchange for all the convenience it offers. While your feelings may vary on how one-sided this relationship is, it's a transaction between us and a massively powerful tech company. Are we supposed to say pretty please for this?
Assistant acquiring the trappings of personhood, speaking to us in the seductive voice of John Legend or whatever your favorite celebrity will be, leaves us in some sense vulnerable to emotional manipulation by an algorithm that knows us very well, but has no emotions of its own.
I've spoken to startups working on voice replication, and it's known in the industry that debt collectors, call centers, and other businesses desire the ability to replicate only the most persuasive voices for use with bots, because when something sounds warm, human, and charismatic, our instincts allow us to be more easily swayed. As we approaching interacting with Assistant as if it were a person, we risk losing sight of the machine behind the mask.
When you think of the children, it's already a little disturbing that by repeating 'Google', over and over again at Assistant (even if now, you only need to do this when starting a conversation), they're being indoctrinated to love and trust a brand while their minds are vulnerable. On top of that, the need to show politeness and deference seems a step too far.
Perhaps if an Assistant speaks with a human voice, it also needs to be treated as a human being. Especially by kids, who might not get the intricacies of what's the difference between talking to a human and talking to an AI. The manners they learn to use with Assistant could carry over into the wider world of human interaction.
But for parents, before you jump at the chance to give Assistant the role of socializing your children, you might consider that it's worth teaching them this difference from an early age, to keep in mind for the connected future they will grow up in.
What do you think? Is it only good manners to treat voice assistants like people? Or should we keep a clear distinction?