The Intersection of Creativity and Artificial Intelligence

Source: Artificial Intelligence on Medium

The Intersection of Creativity and Artificial Intelligence

What is creativity? What is artificial intelligence? And what do they have to do with each other? Take my hand and follow me in the flourishing intersection of artificial intelligence and creativity.

Transformation of a landscape into the style of Monet. Transformed by an AI. Photo by Julien Lavallée.

Artificial intelligence and creativity are both somewhat vague words and often used in a sloppy way. Therefore, let me give you my own definition of both terms before we discover the deep relationship between them.

What is Artificial Intelligence?

AI is probably the buzzword of the century. Everyone is using it. It sells. But in order to grasp what it really means deeply, we have to understand the meaning of intelligence itself thoroughly.

The Oxford Dictionary defines intelligence as “the ability to learn, understand and think in a logical way about things”.

I do not believe that thinking in a logical way about things is key to intelligence, because logic is not soft enough. I’ll come back to that later. Although I believe that the terms to learn and to understand are equivalent, I agree with the first part of the definition with my heart and soul.
Let me elaborate on the equivalence of learning and understanding. Learning and understanding are all about abstraction and compression. Abstraction comes from the Latin word “abstrahere”, which means “to pull awayand is about pulling the noise away from the signal. Abstraction is the process of separating the universal, simple and valuable knowledge from the narrow, complex and unimportant details. And what do we get if we dig gold out of the dirt? Less mass but more value. We compress the value so that we end up with a higher value density. Learning and understanding are both about abstraction. They are about the same thing. They are equivalent.

Intelligence is the ability to learn.

Now that we have defined intelligence as the ability to learn, we have to define what learning is. We have already stated that learning is about abstraction. But where do we find universal, simple and valuable knowledge?

If I want you to remember only one thing from this article then that there are only two things in the universe. Objects and relationships between them. Knowledge lies in these relationships.

In Mathematics the objects are numbers, matrices, functions, etc. and the relationships are operations like addition, matrix multiplication or differentiation. The true knowledge is not in the objects. It is in relationships like the Pythagorean Theorem. In Economy the objects are people or companies and the relationships are transactions of goods, assets, money, etc..
In Physics the objects are physical properties like mass or energy and the relationships are equations like e=mc² .

The history of science is a history of abstraction and compression. At first, there were numerous relationships — expressed as equations — for numerous physical phenomena. Nowadays all equations, that explain the whole universe, fit on a sheet of paper.

It is time to take a deep breath and compress everything we’ve just said into a definition of learning.

Learning is the process of abstracting relationships from data.

I am overjoyed that you have followed me up to here. There is only one missing piece in the puzzle for us to come up with a complete definition of intelligence and to bridge the gap to creativity.

Mathematics is about truth. The way math discovers truth is by proving statements. Computers are about truth as well. They think in zeros and ones, in true or false. Learning is different. Learning is about probabilities. When you learned to ride a bike, you certainly did not prove the optimal actions. You intuitively did the actions with the highest chance of success, while dealing with uncertainties. It feels like (machine) learning is soft and computers are hard. Or to say it with Yann LeCun’s words “machine learning is the science of sloppiness”. Mathematically speaking it is stochastic, not deterministic.

Finally, we can summarize everything we have just said in a definition of intelligence.

Intelligence is the ability to abstract stochastic relationships from data.

Given the definition of intelligence, it is easy to define artificial intelligence.
Every intelligence, except that of natural beings, is artificial intelligence.

What is Creativity?

Creativity is a fundamental desire of human beings. We evolved from the apes and suddenly started creating cave paintings, playing music and telling stories. We build houses, bridges and streets. We invented the wheel, the bicycle and the car. But even though we know what creativity means to us, it is hard to define. To be fair, let us start with a definition of the Cambridge Dictionary this time.

The Cambridge Dictionary defines creativity as “the ability to produce or use original and unusual ideas”.

I do not believe that creativity stops at ideas. It is much broader. Thomas Edison — one of the most creative minds — did not stop at the idea. He built the light bulb. Creativity is also about manifesting an idea in a material object.

Even though I think that original and unusual are synonymous in this context, I absolutely agree with the point that the material or immaterial object has to be original. The real value is in the intersection of truth, opposition and innovation, but it’s easy to confuse innovation and imitation. However, creating something new does not mean that every ingredient has to be new. There is a famous quote by Pablo Picasso “Good Artists Copy; Great Artists Steal”. But be careful, stealing alone is not enough. The art is to reassemble the stolen so that something new and great arises. Innovation happens at the intersections. Hence, I totally agree with the part of the definition that states using original and unusual ideas is part of creativity.

Above all, the process of creation is the most important. And it is inherently stochastic. Let me demonstrate that to you. Assume you would live in a world of zeros and ones, of true and false and would have to paint a painting. There are no probabilities, no uncertainties. In order to take the brush in your hand and paint the first stroke, you would have to prove that this stroke is the optimal action. But this is impossible and therefore you wouldn’t even paint a single stroke. You would stare on a blank canvas forever. Creativity is about probabilities, not truth and perfection kills creativity.

We made it! Now we just have to summarize everything and rephrase the definition of creativity.

Creativity is the ability to create material or immaterial objects by originally combining existing ones while dealing with uncertainties.

Transformation into the style of Picasso. Transformed by an AI. Photo by Nathan Dumlao.

What do they have to do with each other?

Intelligence and creativity are deeply entangled. In evolution, creativity and intelligence emerged in parallel. Humans are not only the smartest, but also the most creative creatures on earth. Albert Einstein once said, “creativity is intelligence having fun”.

Now that we have a definition of both terms, we can explore the deep relationship between them. Perhaps you have already discovered some of the parallels. The most obvious one is that creativity and intelligence are both abilities. Some people are more creative than others and some people are more intelligent than others. However, because it is an ability you can learn it. You can learn to abstract. You can learn to create.

Intelligence is the ability to abstract. Creativity is the ability to create. This sounds similar and it actually is. Abstraction is a special kind of creation.
A kind of creation where the product is smaller, but has more value than the ingredients.

The products of intelligence are stochastic relationships. The products of creativity are material or immaterial objects. It’s the same game again. Stochastic relationships are a special kind of immaterial objects. The relationships between objects are immaterial objects itself.

Additionally, both of them are stochastic. Creativity is about dealing with uncertainties. Intelligence is about stochastic relationships.

Finally, we have found an answer to our question. You probably already guessed it.

Artificial intelligence is a special kind of creativity.
A kind of creativity where the process of creation is a process of abstraction and where the objects are stochastic relationships.