A.I. will come to have a prominent role in society in the next 10 years, but it is not the A.I. that many of us had envisioned.

AIs will play an important role online, masquerading as humans, helping to provide customer assistance and support. Walk into an online store and you are going to be met by an attractive, well-spoken individual. They are going to be knowledgeable, friendly, and helpful. But, they are not going to be real.

AIs will also be used in a number of other ways. For companionship or entertainment, for instance. An interesting, if not disturbing, development.

What is the AI of the future going to look like, and how is it going to work? The AI of the future is going to look a lot like us. It will be an interface, with some sophisticated programming under the hood.

***

The term AI is generally used to describe computers that think, act, or look much like us. In particular, neurologists use the term to describe computers that mimic human neurological systems; computer scientists to describe computers that act like humans - that are able to learn, to exhibit goal oriented behaviors, etc.; and science fiction writers often use it to describe the birth of machine consciousness.

But, for the rest of us, if your computer looks like you and starts talking to you, it is an AI. And the emphasis is not upon what is going on behind the curtain or under the hood (the unseen), but upon the interface (the world of appearance). Thus, the term “Artificial Intelligence” misses the point. It would be more appropriate to use something along the lines of "Artful Impersonator," or "Anthropological Impersonator."

But what is going to make this phenomenon so interesting is not how computers will interact with us, but with how we interact with computers.

We have learned how to interact with people, and we cannot just turn it on or off. It is involuntary. We are going to interact with AIs almost as if they were human; because we, simply put, just can't help it.

Humans respond to cues, such as facial expression and voice tone. This is in part rooted in the neurology of the human brain, and, in part, a result of social conditioning. When someone smiles at us, we tend to smile back (particularly if they happen to be beautiful). And if someone snarls at us, we are likely to be edgy and uncomfortable.

How we respond to various cues is perhaps best illustrated on television. Although we know that we are watching actors, we respond to their facial expression, their tone of voice, their posture. We feel their joy, their hope, and their despair. Or we think we do. But, in truth, they have deceived us.

And it is in this way that AIs will deceive us, as well. We will respond and interact with AIs in very human ways, because that is what we are programmed to do.

Returning to the analogy of the actor, as we watch an actor - as we get to know the character being portrayed - we begin to construct a persona based on very superficial information. We begin to anthropomorphize, characterize, empathize, and to form attachments. And it doesn't really matter if we know it is just fiction. And the same will be true of AIs.

The television also helps fulfill many peoples need for companionship. We may thus ask, to what degree will AIs fulfill our need for attention, affection, understanding, loyalty, etc.? And here we may note that our basic need for companionship may be met, to some extent, in very superficial ways. Owning a pet. Meeting people online.

As AIs become more life-like and convincing, people will begin to have more and more “human” relationships with them. No doubt, a disturbing prospect. But, it is hardly surprising. If you look at much of our daily interaction, it is surprising how superficial and shallow it really is. In fact, it is possible to feel loved and understood when we are not, or not loved and not understood when we are.

But, do we really want to be friends with an AI. Probably not. But, to some extent, we may not be able to help it. There is no doubt, of course, that we will not be entirely satisfied with the knowledge that our interaction is unreal (provided we know that we are interacting with an AI, in the first place), and not entirely satisfied with the experience, which may, in many ways, be lacking.

But for some, AIs will play an important role, whether that of a lover, friend, playmate, or pet. And surely no one would miss loneliness and depression, or the short-comings of their relatives or friends.

***

In review, we should expect to see the emergence of AI in the next 10 years, but it will be a different sort of AI. AI will not be understood as an intelligent computer, but a computer that looks intelligent, and that looks and acts like us.

This development will undoubtedly lead to important developments in business and commerce, and in the service industry, in particular. What makes the development of AIs so interesting, however, is that we will interact with AIs in very human ways, being genetically and socially programmed to do so. People are "designed" to respond to various cues, which enables us exert a surprising amount of influence over them (raising a number of important questions in ethics and design). In this way, we will think of AIs as more than a useful interface, but as our friends.

***

As we reflect upon, and marvel at the computer and its ability to project personality, we are drawn to the question, “How is a computer – composed of little more than sand – able to create and project a powerful, life-like persona? How is a computer able to pull it off?”

We construct a personality – a mirage – from superficial data, and this is the answer to our question. It is a mistake to think of an AI merely as a “computer projection.” The computer does not project a persona. The persona is a projection of our own minds.







***
msc 4/20/08