Updated: Jan 16
We can be surprised only so much to find out that encyclopedias of everything seem now to be available.
Our common experience of using web search engines easily lands us on subjects and facts that we neither already knew about ourselves nor were going to create.
Even further, discovering the extent of information already developed about some new-to-us topic is usually somewhere between humbling and intimidating. We may not know any of the people involved but we compare ourselves to them as if we could be one of them and then intuitively sense how much work has already gone into making the topic, plus the fact that we either wouldn’t or couldn’t do it.
But the same awe comes with simply observing any creative person who has mastered some technique in an activity we don’t know how to do ourselves. We’re caught watching the magician do their trick without even using the misdirection usually needed to leave us marveling at the impossibility.
When computers do the searching or execute techniques, we of course aren't shocked that they can do it. We’ve already seen them doing it for decades, and the main difference that we notice now is how much faster they can get through so much more of it.
What really matters to us, instead, is that the computing winds up being able to sort through everything to present us with something we think is either the “right” result or the “preferred” result.
It is the range of our own criteria that actually makes us decide if the computing is or is not “smart”.
That Ticking Sound
Our idea of “smart” usually comes in any one of three flavors.
There’s “thinking”, which means editing through choices to find ones that we can then relate to each other. Call it logic.
There is “explaining”, which means that we can determine and describe how something that already exists DOES work or IS formed. Call it analysis.
And there is “inventing”, which means describing how a new way of organizing a practical object or action WILL work as needed. Call it design.
Those three things – logic, analysis, and design -- are usually somehow combined for any of three objectives.
One is to Propose.
Another is to Prove.
And another is to Predict.
The vast majority of what we consider “smart” behavior is covered by those three things.
And what excites us most about “smart” is that it can get us something that we need or want, on demand.
Brains and Beauty Too
Among the range of needs and demands is a special case – things that we didn’t already know we needed or wanted. And second in interest level is things that we didn’t already know are possible, and that we experience as being relevant.
Artificial intelligence gets its value mainly from how computers can be trained to refer to our examples of logic, analysis and design – and to apply our techniques of logic, analysis and design. By doing that, it possibly generates results (i.e., produces products) that are needed, preferable, and relevant.
The most important part of the idea of artificial intelligence is the relationship between (a.) the products and (b.) their origin in being smart -- rather than just their being in the right place at the right time for us to stumble upon or receive that product. The idea that A.I. can have "originality" is entirely reliant on its being smart. But we don't have or use the term "Artificial Originality" even though we can explain it.
Artifice, of course, is entirely about fabricating something that isn’t already there. Being smart enough to produce it is what we expect intelligence to enable. Intelligence is the "originality".
Now if we consider what it means to have logic, analysis and design within different practices, one of the most intriguing practices is that of art. Putting this in perspective: we want art because of what it provides. It meets a need or a desire, and we expect it to have relevance.
First we identify what it is that artworks provide to us. Once we identify the characteristics of the art product that support or generate preferability and relevance, the means are needed to produce those characteristics.
The logic, analysis and design in art are means of producing that experience of preference or relevance. That is, the artifice is driven by the intelligence.
Rationally, the notion of “A.I. in art” refers to only two things.
One is the notion that "intelligence" itself can be expressed by computing. The proper name for this, however, is “synthetic” intelligence. Can a computer perform logic, analysis, and design? Yes to all three.
And the other reference is to the goal of applying intelligence, which is to produce an artifact that has the qualities we require from art.
Creativity and the Heart of a New Machine
The term "creativity" is loaded with notions of originality, inspiration, uniqueness, and other conditions that distinguish it as being "special" and, really, a manifestation of a certain kind of consciousness that is not shared even by all humans. This makes it somewhere between implausible and insulting to think that a machine would have it. But what A.I. increasingly presents to us is something that we find harder and harder to distinguish from what many people present to us as creativity.
Here is the important thing to recognize: entertaining though that idea of creativity may be, it is dis-informative, and unreliable to say that the computer is “creative” as if that creativity was at all possible without its essential task, production. When we specifically say “production” we are not meaning to intend “creation” in any lofty other sense. And frankly, when we say "creative", despite our interest in its special glow, we are often only talking about production that is being executed at a level hidden from us like the secret of the magician's trick.
That's just not hard for A.I. to do. What matters, instead, is how we need or want to make use of what it does.