The Conversation: AI ranks in the top percentile of creative thinking

THE CONVERSATION Of all the forms of human intellect that artificial intelligence might be expected to emulate, few people would likely put creativity at the top of their list. Creativity is beautifully mysterious and frustratingly fleeting. It defines us as human beings and seemingly challenges the cold logic behind the silicon curtain of machines.

However, the use of AI for creative endeavors is now on the rise.

New AI tools like DALL-E and Midjourney are increasingly part of creative output, and some have started winning awards for their creative output. The growing impact is both social and economic, just to give an example, the potential of artificial intelligence to generate new and creative content is a flashpoint behind the Hollywood writers’ strike.

And if our recent study of AI’s astounding originality is any indication, the emergence of AI-powered creativity, along with examples of both its promise and dangers, is probably only just beginning.

A mix of novelty and utility

When people are more creative, they respond to a need, goal or problem by generating something new, a product or a solution that didn’t exist before.

In this sense, creativity is the act of combining existing resources, ideas, materials, knowledge in a new way that is useful or rewarding. Very often the result of creative thinking is also surprising and leads to something that the creator did not have and perhaps could not foresee.

It could involve an invention, an unexpected punchline for a joke, or a groundbreaking theory in physics. It could be a unique arrangement of notes, tempo, sounds and lyrics that results in a new song.

So, as a creative thinking researcher, I immediately noticed something interesting in the content generated by the latest versions of artificial intelligence, including GPT-4.

When asked to perform tasks that required creative thinking, the novelty and usefulness of the GPT-4 results reminded me of the creative ideas presented by students and colleagues I had worked with as a teacher and entrepreneur.

The ideas were diverse and surprising, yet relevant and useful at the same time. And, when required, quite imaginative.

Consider the following suggestion offered to GPT-4: Suppose all children grow giant one day a week. What would happen? The ideas generated by GPT-4 touched on culture, economics, psychology, politics, interpersonal communication, transportation, recreation and much more, surprising and unique in terms of new connections generated.

This combination of novelty and utility is hard to pull off, as most scientists, artists, writers, musicians, poets, chefs, founders, engineers, and academics can attest to.

Yet AI seemed to do it and do it well.

Put artificial intelligence to the test

With creativity and entrepreneurship researchers Christian Byrge and Christian Gilde, I decided to test the creative abilities of artificial intelligence by submitting it to the Torrance Tests of Creative Thinking, or TTCT.

The TTCT prompts the candidate to engage in the kinds of creativity required for real-life tasks: asking questions, such as being more resourceful or efficient, guessing cause and effect, or improving a product. It might ask the candidate to suggest ways to improve a children’s toy or imagine the consequences of a hypothetical situation, as the example above demonstrates.

The tests aren’t designed to measure historical creativity, which is what some researchers use to describe the transformative brilliance of figures like Mozart and Einstein. Rather, it assesses individuals’ general creative abilities, often referred to as psychological or personal creativity.

In addition to performing the TTCT via GPT-4 eight times, we also administered the test to 24 of our undergraduate students.

All results were evaluated by qualified reviewers at Scholastic Testing Service, a private testing company that provides scores for the TTCT. Little did they know in advance that some of the tests they were going to grade had been completed by AI.

Because Scholastic Testing Service is a private company, it does not share its guidance with the public. This ensured that GPT-4 would not be able to scour the internet for past hints and related answers. Additionally, the company has a database of thousands of tests completed by college students and adults, providing a large additional control group to compare AI scores against.

Our results?

GPT-4 ranked in the top 1% of test takers for the originality of its ideas. From our research, we believe this represents one of the earliest examples of artificial intelligence meeting or exceeding the original human capacity for thought.

In short, we believe AI models like GPT-4 are capable of producing ideas that people see as unexpected, new, and unique. Other researchers are coming to similar conclusions in their research on artificial intelligence and creativity.

Yes, creativity can be assessed

The emerging creative capacity of artificial intelligence is surprising for a number of reasons.

First, many outside the research community continue to believe that creativity cannot be defined, let alone evaluated. Yet the products of novelty and human ingenuity have been enjoyed, bought and sold for thousands of years. And creative work has been defined and valued in fields like psychology since at least the 1950s.

The person, product, process, and print model of creativity, introduced by researcher Mel Rhodes in 1961, was an attempt to classify the myriad ways in which creativity had been understood and evaluated up to that point. Since then, the understanding of creativity has only grown.

Still others are surprised that the term creativity can be applied to non-human entities such as computers. On this point we tend to agree with the cognitive scientist Margaret Boden, who has argued that the question of whether the term creativity should be applied to artificial intelligence is an aphilosophical rather than a scientific one.

The founders of AI foresaw its creative abilities

It is worth noting that in our research we only studied the results of artificial intelligence. We haven’t studied its creative process, which is probably very different from human thought processes, or the environment in which the ideas were generated. And if we had defined creativity as the need of a human person, then we would have had to conclude, by definition, that artificial intelligence cannot be creative.

But regardless of the debate about the definitions of creativity and the creative process, the products generated by the latest versions of artificial intelligence are new and useful. We believe this satisfies the definition of creativity dominant today in the fields of psychology and science.

Furthermore, the creative capabilities of current iterations of AI are not entirely unexpected.

In their now famous proposal for the 1956 Dartmouth Summer Research Project on Artificial Intelligence, the founders of artificial intelligence highlighted their desire to simulate every aspect of learning or any other characteristic of intelligence, including creativity.

In this same proposal, computer scientist Nathaniel Rochester revealed his motivation: How can I make a machine that shows originality in solving problems?

Apparently, the founders of artificial intelligence believed that creativity, including the originality of ideas, was one of the specific forms of human intelligence that machines could emulate.

In my view, the astonishing creativity scores of the GPT-4 and other AI models highlight a more pressing concern: In US schools, to date, there have been very few official curricula and programs that specifically target and nurture human creativity. Development.

In this sense, the creative capacities now being realized by artificial intelligence may provide a Sputnik moment for educators and all those interested in fostering human creative capacities, including those who see creativity as an essential condition of individual, social and social growth. economic.