Artificial intelligence (AI) is showing its potential in the realm of creativity, a trait many wouldn’t expect from machines. While creativity has long been considered uniquely human, AI is now playing a role in creative production through tools like DALL-E and Midjourney. This trend is gaining momentum, and AI’s creative outputs are even winning awards.

Creativity involves generating new solutions or products in response to a need or problem. It combines existing resources in novel, useful, and often surprising ways. AI, such as GPT-4, is demonstrating similar abilities, producing ideas that are both unexpected and relevant when prompted with creative tasks.

To assess AI’s creative capabilities, researchers had GPT-4 take the Torrance Tests of Creative Thinking (TTCT), which measures general creative abilities of individuals. GPT-4 performed exceptionally well, ranking in the top 1% for originality. This suggests that AI is capable of original thinking comparable to humans.

This emergence of AI-based creativity is unexpected to some, as many believed that creativity couldn’t be defined or measured. However, human innovation has been valued and evaluated for centuries. AI’s ability to generate unexpected and unique ideas challenges conventional notions of creativity.

AI’s creative abilities align with the original vision of AI founders, who aimed to simulate all aspects of human intelligence, including creativity. While AI’s creative process differs from humans’, the outputs are novel and useful, fitting the dominant definition of creativity in psychology and science.

These unexpected creativity scores in AI models underscore the need for educational programs that target human creativity development. The realization of AI’s creative abilities could be a catalyst for educators and others to emphasize the importance of nurturing creativity for individual and societal growth.