Brief Encounters with Artificial Intelligence

I’m preoccupied with how artificial intelligence will impact society even though artificial intelligence isn’t very smart or common yet. A couple of recent observations (the type that come with the metaphorical light bulb over the head1, not made with binoculars from a shelter hidden in the shrubbery) brought me to a greater understanding of how artificial intelligence may fit into everyday life.

First encounter: I picked up my phone and thumbed to my favourite cab company. I hadn’t called them in a year and a half, and was surprised to hear an automated attendant answer instead of the woman with the gravelly voice who called me ‘dear’. It said if I wanted to be picked up immediately at the address I was calling from, I could push ‘1’ and it would happen. I did and did and was (want to be picked up at the address, pushed 1, and the cab arrived shortly thereafter). As we cruised towards my destination, I chatted to the driver, telling him that while the call seemed very efficient, I felt odd about it.

What was lacking? Why was having a machine arrange to pick me up, when that was the very reason I’d called the cab company, disconcerting? If the dispatcher had asked if I wanted to be picked up at home, or the usual place, I’d be glad of the personal service.

For me, the missing link was someone who actually cared if the cab came to get me. Someone who wanted to do their job well and understood that if I was going to the train station it was to catch a train, and being late could have a ripple effect: missing the train, then maybe being late for a job interview and therefore mortgage payments. The gravelly-voiced lady may never have given a rat’s ass if I got where I needed to go on time, but generally, people can empathize with the consequences of being late. Yes, a machine can list the potential consequences, but it can’t remember the time it got a flat tire on the way to a wedding rehearsal where it was the best man for a nervous groom who needed his best buddy’s support.

My second encounter is more subtle. I’m knee-deep in business theory, pondering what gives companies like Harley Davidson and Tim Horton’s their competitive advantage. Causal ambiguity is a thing, which I’ve always liked the sound, and the meaning, of. It’s rather similar to the idea that the sum of the parts is less than the whole. Sure you can make motorcycles growl and sputter in an outlaw kind of way but what makes a Harley the motorcycle of choice for everyone from rock stars to the pastor of an upwardly mobile flock? Why is Tim’s a thing half the Canadian population will wait in line for when the competition across the street has the same product (coffee)? Casual ambiguity. Don’t know how – but it works.

The best explanation we have right now is that casual ambiguity is the sum of a vast number of elements, each subtly different than the norm. Will artificial intelligences be able to dissect casual ambiguity into its multitude of components? It’ll be interesting to see what the AIs come up with to explain the appeal of Hello Kitty or the allure of certain notorious socialites.

On the same topic (competitive advantage), but a more mundane level, human nature dictates we develop processes, or rituals. These get passed down through generations, whether it’s a ham recipe or a manufacturing process for golf balls. Scholars of Operations Management know that the best processes morph overtime because every little detail can’t be recorded, so instructions are interpreted and revised. Will AIs eliminate this natural drift, detailing processes with an infinite number of stepwise instructions? Me, I take detailed instructions, read them until I understand the concept and then construct the functional details.

Humans make mistakes. Sometimes the mistakes are good and we discover something new. Sometimes they’re neutral for many cycles and get propagated through the system then someday turn out to be of benefit.

AIs won’t make mistakes, so there will be no drift in the processes they oversee or serendipitous finding of new things. AIs follow instructions with a precision that leaves nothing to the imagination.

So, we humans, with all our squishy emotions and random actions, are useful after all. We care – who can deliver good customer service without caring? And we err. To err is human, not to err – divinely artificial.

1Does that metaphor still work with a compact fluorescent bulb?

Please follow and like us:

Leave a Reply

Your email address will not be published. Required fields are marked *