A terrible novel is still a novel. A painting can leave us completely indifferent and still be a painting. But introduce artificial intelligence into the creative process, and suddenly we find ourselves arguing about whether the result is art at all.
I find that peculiar.
We seem reasonably comfortable separating what something is from whether we consider it any good.
A song can be dreadful music. A badly acted play does not automatically stop being theatre. Someone can produce an astonishingly ugly sculpture without accidentally creating a new category of household furniture.
Yet our willingness to make that distinction becomes much less reliable when we encounter unfamiliar methods of creation.
We begin examining how something was produced, what the creator intended, how much skill it required, and whether a person was sufficiently involved.
Somewhere inside those questions is an assumption about art that we may never have needed to articulate before.
And that is where the discussion becomes interesting.
The trouble with a familiar word
If you asked me to name examples of art, I could do so without much difficulty.
Books, paintings, music, films, theatre, sculpture.
I would probably include games, certain kinds of architecture and some things whose creators might be surprised to discover that I considered them artistic.
The difficulty begins when we try to establish a rule explaining why all those things belong together.
A beautiful painting might be appreciated for its appearance. A novel can be remarkable without being beautiful in any conventional sense. A performance may leave no lasting physical object behind.
A game can involve carefully constructed environments and stories while allowing the person playing to change the experience through their decisions.
A work may exist to entertain us, disturb us, preserve a memory or explore something its creator does not fully understand.
These activities have similarities, but they do not all appear to be attempting the same thing.
Perhaps that is why definitions of art have such an unfortunate habit of becoming less convincing as soon as somebody produces an interesting counterexample.
The argument has been going on for a while
Plato was concerned with the place of poetry and artistic imitation in human life.
In The Republic, he examines how representations can influence what people believe and feel, and whether certain kinds of poetry belong in a well-ordered society.
Art was already connected to questions about truth, morality and the effects of what people encounter.
Aristotle approached dramatic representation differently.
In Poetics, he explored how tragedy works through the representation of human actions, including the emotional experiences associated with pity and fear.
I am intrigued by the distance between those approaches.
One philosopher worries about the consequences of representation. Another examines its possibilities and how its structure produces an experience.
Neither was considering a language model.
They were examining artistic practices within their own cultural and intellectual worlds, and their ideas remain interesting partly because we can still recognize some of the disagreements.
Centuries later, Leo Tolstoy challenged definitions of art based primarily on beauty.
In What Is Art?, he developed an account of art centred on the communication of feeling.
That changes the emphasis again.
Rather than asking chiefly whether an object is beautiful, we begin asking what passes between the person who creates it and the person who encounters it.
Tolstoy’s account also carries substantial moral demands, and I would hesitate to adopt all his conclusions.
But it reveals something important about the wider discussion.
People have disagreed not merely about which works are good. They have disagreed about what makes an activity artistic in the first place.
A definition can conceal a preference
Suppose I say that art must involve technical skill.
That would seem to accommodate many things we admire.
Then somebody presents an ordinary object in a context that transforms how it is understood.
Perhaps artistic skill is no longer the whole story.
We might instead insist on emotional expression.
But what should we make of a carefully constructed work designed to provoke thought rather than communicate a recognizable feeling?
We could turn toward intention.
Yet people sometimes create things for practical purposes that others later encounter as art.
And we can all think of things whose creators very much intended to be artistic but whose results leave us thoroughly unconvinced.
That final observation leads to another difficulty.
A failed attempt at art may still be art.
We should be careful about definitions that quietly require a work to succeed before it can enter the category.
Otherwise, asking whether something is art becomes almost indistinguishable from asking whether we approve of it.
Those are different questions, and confusing them makes both harder to answer.
The particular difficulty of AI
Artificial intelligence complicates these distinctions because it can participate in activities we usually associate with creative people.
A language model can propose scenes, produce dialogue, suggest descriptions and transform existing prose.
The person using it may provide a detailed plan, choose among alternatives, rewrite the result or discover an entirely different direction through the process.
In other circumstances, someone may request a complete story and publish the first output without making meaningful creative decisions about its contents.
Both processes involve artificial intelligence.
I would find it strange to insist that their artistic circumstances are therefore identical.
But saying that they differ does not resolve the question either.
How much human intention is enough? What kind of decision constitutes an artistic contribution? Does a work require someone to have experienced the feelings it appears to express?
And should the methods used to produce it determine its artistic status independently of the finished work?
There are genuine disagreements here.
Some concern artistic creation. Others concern the rights of people whose work was used to train the technology. Still others involve cultural recognition, employment, ownership or the value we place on human craftsmanship.
They deserve to be considered.
What troubles me is how easily they become compressed into a single declaration about whether something is art.
An ethical objection to a production method is not automatically an aesthetic judgment.
A technically impressive result is not automatically evidence of artistic depth.
And calling something art does not require anyone to admire it.
Why I want to examine the question
Writing Max Aura Dungeon with AI has made these distinctions relevant to something I am actually creating.
The books involve a fictional world, characters, deliberate choices, unexpected developments and a considerable amount of revision.
The technology is part of their creation.
I have found that the questions people ask about such work reveal different understandings of what an author contributes.
Some care most about the words. Others care about the imagination behind them, the labour involved or the relationship between human experience and the finished story.
I am interested in those differences.
Not because I expect to establish one definition that will finally settle art for everyone, but because the effort to define it can reveal what we value.
Sometimes we seem to be discussing an object.
Sometimes we are really discussing its creator.
And sometimes we are defending a cultural boundary without having considered why it was drawn where it was.
I would rather examine those boundaries than pretend they have always been obvious.
If someone believes AI-assisted literature cannot be art, I am interested in understanding the requirement they think it fails to meet.
If someone believes every generated image or sentence automatically qualifies, I would like to know what makes that definition useful.
The answers may be contradictory.
That does not make the questions pointless.
It makes them worth taking seriously.