Can Artificial Intelligence Really Be Creative?
Exploring the Relationship Between AI and Creativity
Can AI really be creative? That’s the million-dollar question.
Given the astonishing rise of AI tools that generate images and texts of all genres and formats in 2022, many people wonder if artificial intelligence can compete with human creativity. In fact, some people fear that AI will destroy human creativity altogether.
Well, the answer to the first question is pretty simple: AI does compete with human creativity.
Creative workers are already becoming obsolete in some areas where it is simply more efficient to use powerful AI-based cognitive tools for ideation processes, or the creation of texts and images. For example, copywriters and illustrators already have to compete with AI tools that do their jobs more efficiently.
Does that mean all copywriters and illustrators will be obsolete? Probably not. But those not using cognitive tools for some of their creative groundwork (e.g. outlining texts or sketching ideas for standardized genres or formats) will have a very hard time staying in demand.
But does that mean artificial intelligence itself is creative? Let’s have a look at what we know about what creativity actually is.
Can Artificial Intelligence Really Be Creative?
Two Types Of Creativity
Scientific research on creativity lives up to its name, having come up with dozens of different definitions for creativity in the last decades. Although true that there are many aspects that come into play when we try to understand what creativity is, in our case, we can use a simple differentiation first:
- Problem-solving creativity involves the ability to find innovative solutions to existing problems.
- Artistic creativity is a form of creative thinking that involves developing new ideas or concepts and communicating them via artistic expression
AI And Problem-Solving Creativity
As the Latin root “creare” for “to create, to generate, to choose” already indicates, problem-solving creativity is the ability to find a solution to a problem by making something. In addition, problem-solving often involves the category of time: getting somewhere during rush hour while your bus line is suspended wouldn’t be very demanding without the implication of “getting somewhere in time”.
So the formula would be:
Problem-solving creativity = coming up with a solution to a problem that
- happens in time
- uses unusual means (originality)
- has been in some way unthinkable prior to the problem-solving
Actually, problem-solving creativity is such an important and basic skill that you can observe it in a lot of organisms.
Here are some examples:
Crows coming up with new ways to crack nuts by leaving them on the road for cars to run over.
Rats coming up with new techniques to reach food in previously unknown habitats and teaching it to their young — and there are even some scientists arguing that microbes swapping genes show some form of problem-solving creativity.
Now what about artificial intelligence?
Unlike crows and rats, which develop creative solutions to problems because they have an interest in surviving, an AI exists in an enviably untroubled and mindless state. Without humans delivering problems, it could not even showcase its superhuman capacity for problem-solving creativity in Healthcare, Law, Finance, Retail, Transportation and Manufatcuring (as Great Learning describes in their post).
Although artificial intelligence does not act consciously or intensionally (yet), it can come up with novel solutions to problem-solving tasks in multiple ways:
- AI-driven problem-solving algorithms are able to examine data and identify patterns,
- which they can then use to generate new and innovative solutions and
- additionally, those algorithms are able to learn from their mistakes and continually improve their performance in order to generate better and more efficient solutions.
Within the above definition, AI meets all the requirements to be certified creativity in problem-solving:
- happens in time ✅
- uses unusual means (originality) ✅
- was in some way unthinkable before the problem was solved ✅
AI and Artistic Creativity
When we talk about artistic creativity, the above definition falls short.
For example, the Mona Lisa would fail to be an expression of a creativity, because its creation did not take place in a short time and by using unknown means, as the above definition demanded:
The sfumato technique was already known, it took about 15 years — some even claim that it was not finished at all — and the “golden ratio” has already been described by Euclid in 300 BC and used in art and architecture ever since.
A Brief History of Artistic Creativity
But if we look at the corresponding artistic period, the Renaissance, we notice that it was characterized by an increased interest in the study of human anatomy, perspective, and idealized forms. Renaissance art is also characterized by its emphasis on naturalism and its interest in classical art and culture. That included the definition of artistic success which in ancient Greece has been measured by
- technical skills of the artist and
- compliance with a certain style’s set of rules
In fact, in ancient Greece, art was a form of discovery in which muses and demons could contribute by mediating inspiration from the Gods, but it was not a creation that owed itself to the individuality of a human being. Creativity was something external, something outside the individual.
This concept was continued in the Judeo-Christian tradition, in which humans were thought to have the ability to create something new only as an expression of God’s work because as a matter of fact, the great magician in the sky had to stay the only being capable of creating.
Interestingly, it was during the Renaissance, that the belief that individual creation was a channel of the divine was challenged and the development of the modern concept of creativity began: creativity would now stem from the abilities of the individual rather than from God. With an intensely human-centric outlook on the world, valuing the achievement of the individual, the Renaissance laid the intellectual groundwork to link the concept of creativity to human imagination.
Back to AI: Does it show Artistic Creativity?
Of course, human and machine cognition are two phenomena too different to apply the notion of “imagination” to both.
However, if we want to tie the concept of creativity to human imagination, one way to attribute artistic creativity to AI could be found by following the claim that the human ego — the thing that thinks it’s imagining things — is only a fabrication of the brain and can therefore be taken out of the equation altogether (a position argued for in the Philosophy of the Mind): neurobiologically it can be shown that the creative process takes place in a dynamic interplay between coherence and incoherence that leads to new and usable neuronal networks. With an ego out of the way, the concept of imagination could be reduced to the discovery of specific neuronal networks that even might have different material bases in machines and humans but could both be functionally the same.
On the other hand, with the human ego still in place, we could look at the so-called Wallas stage model, which explains creative insights and enlightenment from the perspective of the ego through a process consisting of five stages, and try to evaluate where artificial intelligence is actually contributing:
- preparation (preparatory work on a problem that focuses the individual’s mind on the problem and explores the problem’s dimensions)
- incubation (where the problem is internalized into the unconscious mind and nothing appears externally to be happening),
- intimation (the creative person gets a “feeling” that a solution is on its way),
- illumination or insight (where the creative idea bursts forth from its preconscious processing into conscious awareness);
- verification (where the idea is consciously verified, elaborated, and then applied)
In this conceptual framework, artificial intelligence could serve as a creative tool in all the mentioned stages. From the perspective of the collaborating AI, those stages would consist of:
- data collection, 2. data processing, 3. + 4. curation, and 5. application
However, the autonomy with which AI tools operate would lead one to question whether terms such as “co-creation” or “collaborative creativity” would be the most appropriate and meaningful here (in fact, a concept that’s been around since the 1960s, when algorithms began to operate autonomously in so-called “generative art” or “computer art” for the first time).
No matter what philosophical standpoint you want to take here: interesting times ahead, with or without the human ego. While AI technology will increasingly be used as a tool to help human creatives develop new and innovative ideas, it is far from replacing human creativity, which is, after all, embedded in sociocultural meaning games that are so multi-layered an artificial intelligence cannot fully understand or simulate them.
That being said, we also thought that AI would never beat us at the game of Go … 🤔
