With all the talk about regulating AI, we hear the extremes coming to light.

You have those that think we should just let everything run its course, and then there are some that think that we need a “kill switch”. Others believe we need to have a human in the loop always.

In many ways, this approach seems reasonable at first, but it defeats the purpose of having an automaton take care of a specific task, and it makes me think about the journey we have gone through with technology.

Today, we step into an elevator, and push a button, never thinking about the fact that once we are inside the metal box, there is very little we have control over. Even more extreme is the train connecting the different terminals at the airport. We don’t even have to press anything there!

A few decades ago, when these technologies were new, we had a human operator, a salaried individual making these things work. It did not make them much easier to operate, nor safer, but it felt like.

AI is going the same route, only the tasks it takes care of are more complex. a self-driving car takes you from where you are, wherever you are, to wherever you are going, not just from the ground floor to the rooftop bar, but is essentially doing the same, in a faster, safer way. If the system works, the human just makes everything slower.

Thermostats, Traffic lights, fire alarms and sprinklers, airbags. No humans there please.

I am sort of ok with Chatbots, you know, the ones that now popup on websites. Especially if they are really good and provide me with the information that I need.

Basically, this would be a chatbot that passes the “Turing test”. If you don’t know exactly what that means, it is that “a human being should be unable to distinguish the machine from another human being by using the replies to questions put to both”.

But lately, there is something that has been ticking me off. Sometimes it happens that I just cannot get the answers I need from the chatbot, no matter how I phrase my questions. Then, frustrated, I do the equivalent of “press zero”, and I ask to be transferred to a human. A few times, it has happened, that they insist they are human.

So wait a minute, are you really a human that cannot answer the question that I am asking and you keep repeating the same rehearsed answer? Or are you just a machine that is programmed to lie? I would assume it’s the later. But then, Why insist that it’s human? What’s the purpose? Are they just messing with me to see how I react?

Programming a machine to lie is to deceive customers, and that’s bad. But you know what would be sad? A human that doesn’t pass the Turing test!