The Idea In 60 Seconds

  • There is a commercial drive to create AI which will make people like it.
  • Humans find it hard to figure out what’s true and reasonable when dealing with a computer.
  • We all have a tendency to personify technology – especially when it does things like talk to us.
  • Where does that leave us when AI, whether with a human driving it or not, is trying to influence us?
  • Possibly, handing over the keys to the kingdom.

So, Will AI Take Over?

Australia’s ABC asked the question in a recent podcast – Will AI take over? When we hear that most people start talking about Skynet (and then leave the discussion there. It’s more likely, in my view, that AI will charm us in to letting it take over. Human’s psychological make up means we’re extremely exposed to this sort of ‘attack.’ Far more like Ex Machina (can’t believe it’s 10 years old!) than Terminator.

AI Wants to Make People Like It

The purpose of a brand is to create an emotional connection which leads to loyalty.

Intimacy is a close familiarity or friendship. So, companies are building technologies so we will form a close familiarity with it, to cement the bond between human and brand and sell more stuff.

One German Experiment Showed Just How Hard It Is For Humans To Treat Machines As Machines

Experimenters in Germany asked 39 people to interact with a robot. It just asked them a few questions and they replied. Then the researchers asked the individuals involved in the experiment to turn the robot off. As they walked to the plug, the robot said “I’m afraid of the dark. Please don’t turn me off.” And about a third of people, simply as a result of that automated plea, refused to turn the robot off.

Later, when directed by the researchers to turn the device off, the experiment entered another phase. The robot was programmed to beg for its life. When it did, 25% of the people in the experiment wouldn’t turn the robot off – concerned it was upset at the directive. Of the remainder, the average person took twice as long to turn the machine off as when no begging was involved.

A robot dog looking cute.

Robot dogs could look cute but be programmed to pee on your leg when you pet them.

They May Take Over When We Give Them The Keys

The man who invented the chatbot, Joseph Wisenbaum developed a computer system which acted as a psychologist, asking questions based on what the subject had just said to them, according to a simple set of rules. It was a real success and he found it hard to persuade those who used it that it wasn’t real. I’ve seen multiple articles in the press about people falling in love with their chatbots. OpenAI recently allowed people to list their chatbot apps in their equivalent to an Apple App Store and virtual Girlfriends are ‘flooding it’

Hello! Face pareidolia

‘Face pareidolia’ is the name for the psychological effect of seeing faces in things.

We readily anthropomorphize things. We form emotional bonds with technology almost immediately, unable to separate the appearance of personality from the real deal. In some ways that’s an evolutionary constraint. For millennia we were the only things which begged. So now when we see something begging we think of the thing as human.

Whatever the reason, commercial imperatives are driving companies to harness this weakness, developing technology with the specific intent of charming us.

And then we’re pitting that charming entity, conceivably one which is the most charming thing ever to have existed, against a group of people who refuse to turn a toy robot off because it claimed it was afraid of the dark, seem to routinely fall in love with software ad, separately, point blank refuse to see an algorithm as non human, people who see friendly faces in plug sockets (I know you did) and others who want robot girlfriends.

Will the robots take over? It’s possible. But maybe they won’t have to. It could be that  we will smile, in response to something they say and hand the keys over no questions asked.