I recently spoke with a visual artist and heard her explain how a passing cloud shadow completely changed her idea for a painting. She felt a hunch, but no rational reason. It’s difficult to picture a machine having that kind of instinct.
Artificial intelligence has significantly increased its footprint in creative fields during the last several years. Machines are already quickly becoming into what many refer to as “co-creators,” producing anything from poetry to logo design. Predicting creativity, the essence of humanity, is a completely different matter.
Key Context –Can a Supercomputer Predict Human Creativity?
| Element | Description |
|---|---|
| Central Question | Whether machines can meaningfully predict or replicate human creativity |
| Key Technology | AI and large language models like GPT-4 |
| Human Advantage | Emotions, intentionality, intuition, and lived experience |
| AI Strength | Pattern recognition, divergent associations, high-speed content generation |
| Notable Test Used | Divergent Association Task (DAT) developed by Jay Olson |
| Surprising Finding | Some AI models outperformed average human scores on DAT |
| Limitation of AI | Lacks subjective experience and motivation |
| Study Source |
Today’s supercomputers replicate artistic outputs at a scale that was previously unthinkable by utilizing enormous datasets. Their language models provide engaging answers to prompts because they have been trained on a wide range of materials, including scientific publications and literature from the 19th century. They write short stories, create tunes, and paint abstract scenes. However, something essential is lacking.
Intention is at the core of human creativity. Not merely a response to facts, but a desire influenced by feelings, memories, and even inconsistencies. Longing is frequently at the heart of love poetry. There is typically a hidden narrative when a dancer improvises. These are expressions rather than only outputs. Despite their intelligence, supercomputers are unable to communicate themselves. They do calculations.
The Divergent Association Task (DAT), created by Jay Olson, is one particularly creative exam that has drawn interest for its efforts to gauge divergent thinking. The DAT asks participants to respond to a cue, such as “cooking,” with ten unrelated words. Words like “pan” and “stove” are examples of linear thinking. However, responses like “mirage” or “oxygen” delve farther into unanticipated territory.
Remarkably, when researchers tested a number of sophisticated AI models, such as GPT-4, they discovered that the models performed better than the typical person across tens of thousands of responses. That caused a stir in the fields of computers and psychology.
The results first gave the impression that AI would be able to “think outside the box.” However, when there isn’t even a box—no internal experience, no sense of boundaries—what does it actually mean? Conventions are not broken by machines. They use probability to repurpose them.
However, humans frequently experience discomfort or rebellion when they defy convention. Frustration, excitement, anxiety, or even protest can all inspire creativity. Only because a comparable phrase was previously in its training set can a supercomputer come up with a breakthrough slogan. Outrage wasn’t created by it. It repeated it.
Olson and his team documented how artificial models imitate human responses by working with researchers from the UNIQUE Quebec Neuro-AI center. Nevertheless, the study warned against relying too much on measurements. After all, it’s hard to define creativity. Tests can score results, but not processes. And humanity shines in the process.
As I looked over the dataset, I stopped at that thought. Although the AI results were emotionally flat, they were quantitatively excellent. That contrast served as a reminder to me that creativity is more than just novelty; it is context-infused meaning.
There is no preexisting template when a child names a marshmallow city or draws wings on a cat. Their works are individualized and frequently motivated by feelings or whims that a computer cannot detect. AI finds it difficult to replicate, let alone forecast, the subjectivity and randomness.
Because of this, a lot of artists now see AI as a tool rather than a competitor—a highly adaptable assistant that simplifies tedious tasks and stimulates creativity. It can mimic brushstroke styles, offer color schemes, or even suggest rhymes. Still, a human hand is needed to determine what is valuable.
Prediction becomes less of a prophetic ability and more of a probability engine in this situation. AI may be able to choose the best story structure for a screenwriting competition. However, it is unable to foresee when a writer may decide to completely eschew narrative conventions and produce something unvarnished, unadulterated, and entirely personal.
Researchers have improved the fluidity and stylistic adaptability of AI models through strategic modeling. However, flexibility does not equate to creativity. A creative twist doesn’t make AI feel proud. Before publishing a song, it doesn’t suffer with doubt. Only input and output are present; there is no internal journey.
However, it’s crucial to remember that there are situations where it’s difficult to distinguish between human and machine creativity. For instance, musicians may incorporate AI-generated loops into a song and then add improvisation. Sometimes, graphic artists employ font that is provided by a machine and refine it instinctively. These hybrid methods are becoming increasingly popular and frequently produce very successful outcomes.
The effect has proven especially helpful for medium-sized creative teams. AI frees up time for more in-depth work by automating drafts and prototypes. “AI writes the copy, we tell the story” was written on a whiteboard at one advertising agency I recently visited. It was hopeful, almost cooperative.
Supercomputers will probably be used in creative work much more in the years to come. They will make suggestions, remix, then remix once again. However, it will be difficult to forecast creativity in its purest form, where creation is spurred by an unsolved memory, a passing cloud, or a restless night.
Because true creativity isn’t limited to problem-solving. It investigates one. It lingers, contemplates, and sometimes contradicts itself. Efficiency is not its goal. It seeks to express what only a human being, with all of their lovely flaws, could.