Siri, would you set an alarm for 3:30 please?
Siri, set an alarm for 3:30.
Damn it, Siri, I told you to set an alarm for 3:30
As you know, “Siri” is a “personal voice search assistant.” She (Siri has a female voice) is a feature of the iPhones and she doesn’t care how you address her. She has no preference among the three “requests” above. Unfortunately, I do. Let’s look at that. I believe this is the woman who voices Siri. She looks like a pleasant person.
One of the hundreds of gender-related experiments cited in Nancy Etcoff’s, The Survival of the Prettiest is one where the men who are the subject of this experiment are given a picture of the woman they are talking to on the phone. It’s the same woman in every case, reading the same script in every case, in the same manner in every case. The difference—the “variable” we social scientists like to say, hoping that someone will mistake us for physicists—is the picture of the woman. A picture of a very attractive woman is given to some of the men; a picture of a very plain woman to the others. You can tell by listening to a recording of the men which woman they think they are talking to.
Why? Well, like everyone else, these men have learned to tailor their manner to the audience. If you do that long enough with enough audiences, you learn quite a few styles of presenting yourself. And if you accept the notion of “authentic character,” (rather than situational role playing) you come to an expectation of what “the real you” sounds like. And also, what you ought to sound like.
This is all pretty complicated for human-to-human interactions, but today, we are considering human-to-operating system “interactions.” Since Siri doesn’t actually care how I address her, I need only care about how I feel, talking to her one way or the other. I can treat her like an invaluable personal assistant. I can treat her like a colleague. I can treat her as the drone at the office that I have to work with no matter how much I wish she would call in sick more. None of those affect her. All of them affect me. This, by the way, is a picture of Siri Hustvedt, author of “The Blazing World.” She comes up when you Google “Siri,” and I had no idea why.
So far, I have chosen collegiality. I’ve tried being more polite, hoping that I would feel better about myself that way. It doesn’t work. I feel stupid. I imagine that someone is going to tap me on the shoulder and say, “She’s not really in there.” So I compromise: I say, “Set an alarm for 3:30.”
I’m one of these guys who speaks differently to women than I do to men. It was the height of the “women’s lib” season when I was in grad school in the early 1970s and I got a lot of grief for doing that. The “liberation norm” of that season was that treating males differently from females in any form of social interaction was a form of social control. I didn’t like the idea of treating men and women alike, and I had been practicing the two styles for several decades before I ran into this reaction and I didn’t like getting scolded. So, all in all, I didn’t change then and I haven’t changed in that way since.
This brings me to an issue with Siri, who sends all the “woman” signals I learned to respond to in a gendered way. There is no reason, of course, that I couldn’t have a male voice on my phone. Here’s a little fantasy I found somewhere this morning.
I like Mark. He wakes me up at six on Mondays, Wednesdays and Fridays, knows not to disturb me on Thursdays. Often, he is the reason I remember birthdays, anniversaries and appointments with doctors. He even calls people for me when I’m ill or just plain lazy. So, a few months after getting to know him, I asked, “Mark do you like me?” “I would not wish any companion in the world but you,” he said.
Mark is the “interactive personal search assistant”—the Siri—on her phone. It seems to me that she isn’t having the kinds of difficulties I am having. So one of the options is to embrace the possibilities that the interaction opens up. Men and women could have “significant other” voices put on their phones. We may not be far away from having the actual voice of the actual “significant other” recorded, “Sirified,” and used on our phones. “Bette,” I would say to my iPhone, “would you wake me up at 3:30?” My actual Bette could be in Heidelberg but she’s still kind of here “in the person of” my operating system.
Or you could go the other way. You could have the voice of Rajib, the house boy. You could have the voice of Rastus, the house…um…person, from a pre-Civil War movie epic. Then you could order these people to do whatever today’s Siri does and do it in the “I am your colonial master” voice or in your “I am your owner and don’t you forget it” voice. Picture that interaction on a crowded public transit system.
The point here is that the operating system doesn’t care and you are going to have to consult “who you are willing to seem to be” in order to choose a style. Sherry Turkle’s research with human/robot interactions shows that mostly, we want them to like us. We want them to succeed. We wish them well. He hope they feel as positively toward us as we do toward them.
And that’s just the beginning. It is, technologically, a piece of cake to set the desired “style” of interaction of the operating system. Who “desires” this style? Not an easy question to answer. It is easy to say that the programmers “desire” the style of whatever they devise and put in the operating system and that is true in a very limited sense. They did write the code. But they did not choose the style. The marketing department—someone with an advanced degree in cognitive neuroscience—chose the style.
Having “a preferred style” means that if you speak more brusquely to the operating system than she is set for—imagine a thermostat-like “setting”—she takes offense. Or she wonders if you are having a bad day. She introduces a trace of reluctance or of hurt feelings into her statement that the task you ordered is done. She says, “I have set an alarm for 3:30” in a way that suggest she would rather not have done it.
I don’t suppose I will have to solve this problem myself and I am sure that my grandchildren will find a way. Probably, I should just leave them to it. I would like to come up with a solution in principle, though.
“Siri, is there any good way to handle this dilemma?”
“I’m sorry. I don’t understand the question.”
“Never mind. Probably I don’t either.”
 As you see, you want to start dropping quotation marks everywhere so you don’t fall into an anthromorphic maze with no hope of escape. We are already in the maze. There IS no hope of escape. I’m going to stop with the quotation marks, right after this. Siri is a “she,” who “lives” in my phone and who “interacts” with me.
 To be fair, this little scenario is presented as “what the future of interaction with operating systems is going.” There actually is, however, a male voice called Mark and he is now available.
 Alone Together. The best book about human/robot interactions I have read.