The Revolution of Computational Creativity
When the finance company JPMorgan Chase announced that it was replacing human copywriters with an algorithm, some thought it was a marketing ploy riding the crest of the AI wave. However, those who knew something about so-called artificial intelligence knew that this was just the beginning. Computational creativity was breaking from academia into real-life applications in the creative industries.
Chase contrasted the man and the machine to find out if machines had anything to give in copywriting. As it turned out, the marketing messages written by the algorithm were delightful for Chase and depressing for human creatives who took an ugly beating on click-through rates. So Chase looked at the data and switched to using algorithms instead of humans.
Stunning examples of how machines wedge into the creative field alongside and in some cases without much help from humans have since surged: machines compose music, create works of art, design product names, create brand identities, make recipes, and predict protein shapes in drug development.
However, the onslaught of similar examples is yet to come. We are transforming the world into mathematics at such a pace that we will soon hardly remember how product development, marketing, and entertainment were ever done by humans alone.
Still, we do not yet fully grasp the profound difference between human and machine creativity, and how that difference will change the creative field.
The creative output of a machine is a prediction. For example, if a machine creates a new recipe, it predicts what kind of recipe would work well based on the data and the restrictions given to the algorithm. In practice, it merges insight, concept production, testing, and selection and transfers it from the physical world to the laptop and the cloud. People design, research, craft understanding, create, compare, and undertake consumer tests and focus groups. In the mechanical process, all of those outcomes are already inside the final output. Computational creativity collapses time, reducing the creative process’s latency from months and years to hours, or even milliseconds in some cases. Machines can iterate millions of product versions, compare and test them in a time that humans can hardly get their research planning meeting started.
There are also qualitative differences between the computational and human process. For example, people do not know the complete picture of the data they have absorbed in the creative process. On the other hand, in computational creativity, we can always go back to the data and test whether our model predicts the phenomenon we aim to predict. Additionally, there is a vast difference between humans and machines in how much data we can analyze within the given timeframe.
The machine can also process far more data than a human and perform countless experiments in a short time. With the machines, we get as much memory capacity and computing power as the company’s credit limit allows, and everything happens so fast after the data is collected and cleaned.
The prospects for the future for computational creativity are dizzying. There is nothing to prevent digital products and services from being created at the very moment a customer opens their interface. So, just as the algorithm predicts which message will work for Chase’s customer, it could predict what kind of product is suitable for the customer at that specific point and time of interaction and create — that is, predict — it on the fly. Thus, computational creativity is accelerating our world into a state of what I call zero latency.
However, there are things where we humans are superior, and machines stink. The machines are helplessly miserable in-between categories and domains. While the machine will soon be able to write — that is, predict — its first Hollywood script, the machine cannot create a new art form or understand the meanings associated with it. It cannot evaluate anything that is outside of the boundaries of a given category. If a machine accidentally produces something inherently original, it must be the human co-creator who can understand the value of that accident. (And yes, you can program more “accidents,” but it is still the human who makes the value judgment. Indeed, the creative role of man is increasingly to look between categories, understand the value, and search for and create meanings. Thus, little by little, human creative work begins to slide more and more into the field of art, to conduct transgressions, and to embody meanings. Yet, paradoxically, it seems the machine encourages us, humans, to reflect on meanings and make art.
I welcome that change in the current environment full of lifeless products and hollow motifs.
I will conclude this article with an instruction written by an algorithm ideated by me and implemented by data-scientist Jarno Kartela: If you are looking for a direction, a path, a frame, a path — take the first step! You don’t want to be a prisoner in the gray area. Find small things you can control. Make them your own. Take the first step and discover what your future can hold.
About the Author:
Henri Hyppönen is Finlabs Chief Innovation Officer and resident inspirational leader. At Finlabs, he leads innovation across the organization to push the limits of our team and to educate and enlighten clients on where to drive their business and products.
Henri speaks regularly about the future, business, leadership & creativity, and about the intersection of human and machine intelligence. He has published five books. The most recent (2020) on the future of creativity.
Henri inspires us every day at Finlabs with his ideas and insights. If you’d like to learn more from Henri and discover how Finlabs can work with you to build intelligent digital products to drive your businesses forward with purpose, drop us a line.