This groundbreaking achievement showcases the impressive capabilities of artificial intelligence, demonstrating its potential to challenge human creativity.
To assess the creative abilities of AI, researchers conducted an experiment using the Torrance Tests of Creative Thinking (TTCT). These tests simulate real-life creative tasks, such as problem-solving, resourcefulness, cause and effect analysis, and product improvement. According to a new study by the University of Montana, the researchers administered the TTCT to both GPT-4 and a group of undergraduate students. The results were then evaluated by trained reviewers at Scholastic Testing Service, a private testing company.
Surprisingly, GPT-4 scored in the top 1% of test-takers for the originality of its ideas. This achievement indicates that AI has the capacity to meet or even exceed human abilities in terms of original thinking. It's important to note that the tests did not measure historical creativity, which refers to the transformative brilliance of figures like Mozart and Einstein. Instead, the focus was on assessing general creative abilities, also known as psychological or personal creativity. The results highlight the potential of AI to generate unexpected, novel, and unique ideas.
The emergence of AI's creative abilities challenges some long-held assumptions. Many people believe that creativity is difficult to define and evaluate. However, throughout history, innovative and imaginative human creations have been recognized, valued, and even traded. Fields like psychology have been scoring creative work since the 1950s. The person, product, process, press model of creativity, introduced by researcher Mel Rhodes in 1961, categorizes the different ways creativity has been understood and evaluated.
Furthermore, there is a debate about whether the term "creativity" can be attributed to non-human entities like computers. Cognitive scientist Margaret Boden suggests that this is a philosophical rather than a scientific question. Regardless of this debate, AI's current iterations are capable of generating novel and useful products. This aligns with the dominant definition of creativity in psychology and science. The founders of AI themselves recognized the potential for machines to exhibit creativity, including the originality of ideas.
While the research focused on the output of AI, it did not delve into the creative process or the environment in which the ideas were generated. The creative process of AI is likely different from human thinking processes, and the context plays a crucial role in shaping creativity. Nevertheless, the fact remains that AI's products demonstrate novelty and utility. These findings should prompt educators and individuals interested in fostering human creative abilities to take note.
The current creative abilities of AI could be seen as a "Sputnik moment". With a lack of official programs and curricula targeting human creativity, the realization of AI's creative potential highlights the need to prioritize and cultivate human creativity. Creativity is essential for individual, social, and economic growth, and AI's capabilities can serve as a catalyst for further exploration and development in this area.