Is AI creative?
A new study conducted by Dr. Erik Guzik from the University of Montana’s College of Business suggests that AI, specifically GPT-4, can exhibit creativity in the same way that humans do. This challenges the notion that AI can only emulate creativity and presents a more complex perspective on the subject. According to Guzik, creativity occurs when individuals combine existing resources in a novel way to generate a new solution or product. When prompted with tasks requiring creative thinking, GPT-4’s output showed novelty, usefulness, and even surprise, similar to the ideas produced by human students and colleagues.
In order to evaluate the creativity of AI, Guzik utilized the Torrance Tests of Creative Thinking (TTCT), which measure fluency, flexibility, originality, and elaboration. When compared to human test-takers, GPT-4 ranked in the top 1% for fluency and originality. This discovery challenges long-held beliefs about the nature of creativity and raises important questions about how we define and assess it.
We may not really understand creativity
While AI’s ability to demonstrate creativity is intriguing, there are still many complexities surrounding the concept. The study’s findings prompt us to reevaluate our understanding of creativity and question whether it can accurately be defined and scored. It also raises the philosophical debate of whether the term “creativity” can be applied to nonhuman entities like computers. Cognitive scientist Margaret Boden argues that this is a philosophical question rather than a scientific one.
In a press conference, Guzik shared an insightful response from GPT-4 when asked about its own performance on the test. The AI acknowledged that our understanding of human creativity may be flawed and suggested the need for more sophisticated assessment tools to differentiate between human and AI-generated ideas.
A Sputnik moment for creativity
This study signifies a pivotal moment for assessing and fostering creativity, particularly in educational settings. Our current approaches may not accurately identify and nurture creative thinking, which could hinder the development of future thinkers, artists, and innovators. The implications of AI’s creative capabilities challenge traditional notions of creativity and call for a more nuanced and effective system.
Guzik sees this as a “Sputnik moment” in the field of studying creativity, referring to the launch of the Sputnik satellite in 1957 that catalyzed investment in science and technology education. He believes that AI’s creative abilities can serve as inspiration to further human creative capacities, ultimately leading to individual, social, and economic growth.
As our understanding of AI and creativity continues to evolve, it is crucial to reassess how we approach and nurture creative thinking. The study conducted by Guzik sheds light on the potential of AI to exhibit creativity and encourages us to explore new possibilities in the intersection of human and artificial creativity.
Journal Reference: Erik E. Guzik et al, The Originality of Machines: AI Takes the Torrance Test., Journal of Creativity (2023). DOI: 10.1016/j.yjoc.2023.100065
Editor Notes
In this groundbreaking study led by Dr. Erik Guzik, the creative abilities of AI have been brought into the spotlight. The findings challenge our preconceived notions about AI’s capacity for creativity and call for a reimagining of how we assess and foster creative thinking. As AI continues to advance, it is essential to stay open-minded and explore the possibilities it presents in various fields, including art, innovation, and education. By embracing AI’s creative potential, we can unlock new avenues for growth and progress.
To stay updated on the latest news and developments in the world of AI, visit GPT News Room.
