Creativity (Only Human?)

Computers Can’t do “AHA!”

Until last month, I thought all creativity was human. I never thought of any other kind. I was made aware of non-human creativity and had to learn about it. This is a shallow dive into non-human and human creativity.

Non-Human Creativity

Computational creativity is an active academic research platform. Researchers from around the world are asking questions about how to program computers like human neural networks to compose music and create visual and language arts. A good place to learn more is the Association of Computational Creativity. Tools for businesses and militaries are being developed as well. Autonomous shuttles with human operators onboard are being evaluated, as are driverless cars.

Human Creativity

Human creativity does not have an agreed-upon definition. “Novel and useful” are the defining words. They imply curiosity and imagination, but they leave much to the imagination about creativity. Here is a simple example of human creativity:

From the culinary world, I once made a creative Thanksgiving dinner. Instead of the traditional roasted turkey, mashed potatoes, sage dressing, roasted winter squash, and lakes of gravy, I made a Thanksgiving pizza. Ingredients were arranged on pizza dough and baked and served with a cranberry sauce side. This certainly created many things. Thanksgiving pizza is now a legend; it is something to be avoided like Medusa. It created a trip to the freezer to grab a pepperoni pizza. It did not create the hoped-for culinary delight that would be enjoyed for many years.

The 4 P’s of creativity, Person, Process, Product, and Press (environment) postulated by Rhodes (1961) remains useful, though a not perfect model for evaluating creativity. I was the person, and I am curious and value originality. I have domain competence, and I can evaluate outcomes. The process was a new recipe. The order of layering ingredients was decided. The product was unique. The environment was favorable, my family welcomed the new pizza, and they were fair in the evaluation of the product. I was open to rejection or hearing ways I might improve on the pizza. All 4 Ps are met — not bad.

Computational Creativity/Co-Creativity

The goal of the Association for Computational Creativity is to construct a program or computer that can perform “human-level creativity”. Other goals include creating algorithms of human creativity to better understand and define human creativity and writing programs to assist humans to create. If the goal is to construct a program or computer, that means it has not yet been achieved. Computers are not creative, but there are many examples of humans using computers to be creative in the shared space of co-creativity.

Humans program computers to create alternatives, and test those options are most likely to succeed based on probabilities. Examples of co-creativity are:

• In the movie War Games, a teenager hacks a military computer and asks to play a game called “Global Thermonuclear War”. The computer, named Joshua, was able to run through all the iterations of strategies that could be used to win this terrifying game. When Joshua realized that the game had no winners, it did not suggest that everyone engage in a new and unique game it designed. Rather, it suggested a nice game of chess that it was also programmed to play.
• AARON is a computer program written by the artist Harold Cohen. AARON can draw and color distinctly different images. The images are based on complex algorithms. New art is created, but it is based on Cohen’s programming, not AARON making independent creative decisions.
• IBM’s Deep Blue supercomputer was programmed to win the game of chess. It lost to World Chess Champion Garry Kasparov, but in a rematch in 1997 Deep Blue bested Kasparov. Deep Blue was able to evaluate hundreds of millions of moves per second and won by evaluating potential moves by the probability of success.

The examples stack up poorly to the 4 Ps as computational creativity. First, they are not persons. They depended on a human to write the programs. The examples are unique and produce useful solutions, like there is no strategy to win a global thermonuclear war, but is that a novel solution? All three of the examples are human-dependent.

As examples of co-creativity fit the 4 Ps well. The programmers are people, they have competence in their intellectual domain, they are evaluative persons. The process they used was new, and the products were unique. The environment for creating human/machine interaction is burgeoning with new ideas for methods computers can make the life of humans better.

Having an “aha” is creating many ideas and mashing them up to a workable product or process, it is an evaluation of the many ideas to find the one that will work. An aha idea must be implemented to complete the creative cycle. Implementation leads to new challenges that require creative solutions. A computer can do those things only if programmed by a human. The person part of the 4 Ps is missing.

Questions

Many questions remain unanswered:

• Can a computer be creative when the definition of creativity remains elusive?
• Intellectual property rights, patents, and copyright are assigned to humans. Can they be assigned to computers?
• If a computer commits a crime, how will it be managed in judicial systems?
• If creativity can be spiritual, can computers be spiritual?
• Is computational creativity a goal we want to achieve?

Don’t worry if you are undecided about any of these questions. The fact that your lights have begun to flicker, and your electronic temperature controls are acting up is almost certainly unrelated to a superior computer…

Rhodes, M. (1961). An analysis of creativity. The Phi Delta Kappan, 42(7), 305–310.