Behaving Intelligently

After Turing, there was a shift in developing artificial intelligence that moved away from trying to produce a thinking machine to trying to produce a machine that behaves as though it thinks. Today these machines are ubiquitous, from the self-parking cars on the street, to the personal assistant in your pocket. And although it’s obvious that machines and humans are not synonymous, the problem now is in the computational model of psychology, which apparently seems so definitive of both agents’ inner-workings.

This model focuses on the input and output of computation. So if a specific environment is going to be perceived then we can count on a specific reaction to that environment based on what we know of its perception. In other words, behavior will always happen within a context. For programming an AI machine there are certain parameters that can be established right away to have it behave under certain circumstances- for example, programming a vacuum that will clean up a room if there’s dirt, and not clean up the room if there’s no dirt- it’s very simple. What isn’t so simple is the limitation surrounding it.

Humans, in a lot of ways, have been reduced to this same kind of simple behavior. There is nothing to distinguish how AI acts and how we act.

Through urban planning, architecture, advertising, and etc. our environments can be constructed in such a way as to elicit certain behaviors- think of the structure of a shopping mall, for instance. We can be taught to perceive the environment in a selective manner, or according to certain parameters, or to whatever constitutes real input. You could argue that this limits our capabilities on a much larger scale. The fact that we think driving, for example, is a human activity and machines can do it, is really sad for us. Or even the fact that people are so responsive to Siri’s programmed sense of humor. What we all do when we think is a process of computation that machines can ultimately do much better than we can, and this seems to be related to our technologization, in general. It’s as if we might as well just have machines and only machines.

Our saving grace is that there is something more fluid at the basis of human intelligence (although I’m not even sure if “intelligence” is the right word for it.) We need to figure out what it means to actually think, and more importantly, we need to start doing it more.