“An iPhone is not a robot. It’s a personal assistant. You should be better acquainted with it.” So scolded a tattooed techie at the Apple Store’s Genius Bar. I gave a guilty shrug but did not say anything. There seemed no point in explaining that my only concept of a personal assistant was a living person. Someone to light and guide me—more like Jeeves than Alexa. I wanted the sound of a voice that came from the diaphragm, not from an algorithm . It made me lonely to think of being left to the aid and companionship of this computer in my purse.
Even doctors seem to be receding behind apps and the distancings of telehealth. As telemedicine moves the doctor-patient relationship out of examining rooms and onto digital platforms will doctors fade further into interchangeable “health care providers”?
Speaking With A Temperamental Robocaller
I had touched something poisonous in the garden: ivy, oak, sumac, a yellow sac spider? After a week of living with a galloping allergic reaction—blisters spreading down to my feet—I took my rash to the dermatologist. It was a speedy visit. Just enough time for her to say, “Well, it’s not infected yet.” Then she turned to the computer to email a script off to my local pharmacy. (No more handwritten scripts.)
Her practice’s robo-caller spent more time with me than she did. Next day it left a message on my answering machine asking me to dial an 800 number and rate my doctor. I shrink from calls like that. They turn a patient—one person under the trusting care of another—into a common consumer of services. I erased the call. But the robot was insistent. It dialed back an hour later, this time asking me to rate, from one to five, my office “experience.” I ignored it. Next evening, at dinner time, the pushy thing rang me again to say, “This call is for Maureen Mullarkey. If you are Maureen Mullarkey, press One. If you are not, press Two.”
I knew what the robot was up to and why it was calling. But now it was my turn to try a question of my own. Bypassing instructions, I asked: “Please, what is the purpose of this call?” The robo-voice did not hesitate for a nanosecond. It shot right back: “That’s an inappropriate response.” And it hung up.
A bit snappish, yes? You would think robots would be programmed with a more flexible adaptive system. That they might be capable of a range of ways to punt: “I know it’s a nuisance. But I can only handle the question asked.” It would be nice to hear, even from a robot: “Please take your time and try again.” Or simply, “I’m sorry. I don’t understand.”
Algorithms Can Be Temperamental
My exchange with this cybergal—the voice was female-ish—illustrated roboticist Rodney Brooks ‘ contention that algorithms do not like people. As he puts it, “They have no idea what a person is.” Despite stating that generalizations applicable to humans do not apply to AI systems, the former director of M.I.T.’s Artificial Intelligence Laboratory, speaks of algorithmic intelligence in terms we do associate with humans. Brooks’ own language grants them temperaments:
Long before there are evil super-intelligences that want to get rid of us, there will be somewhat less intelligent, less belligerent machines. Before that, there will be really grumpy machines. Before that, quite annoying machines. And before them, arrogant, unpleasant machines.
Unsettling, his comment intrigues and disturbs at the same time. By definition, AI is intelligent. But is it reasonable? Is it sensible in the root meaning of the word? Intelligence without introspection or empathy, without guilt, remorse, or a sense of sin—what is that but sociopathology transferred to a machine.
Donald Knuth , world-recognized algorithmist, cautioned against algorithms that are not written by humans, but by the machine as it learns. Yes, programmers still train the machine and feed it data. But there is something more worrisome than bugs and biases in the data. Knuth quotes his colleague Kevin Slavin, a research affiliate at M.I.T.’s Media Lab:
We are now writing algorithms we cannot read. That makes this a unique moment in history, in that we are subject to ideas and actions and efforts by a set of physics that have human origins without human comprehension.
To Return To The Starting Point
This began with a lament that an iPhone is no substitute for a personal assistant. So then, what are such persons made of? When are they on call? What degree of intimacy should they have? This painting explains. Here is Voltaire in his night shirt, putting on his trousers while dictating to his secretary. Callini was dressed, his quill sharpened, and at Voltaire’s bedside as the writer woke up.