Creativity can feel like a flash of inspiration. Research from the University of South Florida suggests it may also be something we can deliberately steer.
Shai Vardi
Shai Vardi, assistant professor of information systems, and co-author Vidyanand Choudhary of the University of California, Irvine, developed a mathematical framework that helps explain how people — and artificial intelligence — move from familiar ideas toward more unexpected ones.
The researchers model creativity as a random walk through a network of ideas. Concepts are points in the network, and the connections represent how likely the person or AI system is to move from one idea to another. Creativity techniques work by changing the path through the network, making less obvious destinations easier to reach.
The framework not only explains why existing creativity techniques work but can also be used to design new ones systematically. The researchers derived new interventions from the model and tested them in creativity tasks involving human participants and multiple large language models. These interventions increased the novelty of responses, although their effectiveness varied across settings.
The findings, published in Information Systems Research, could inform product innovation, advertising, and AI prompt design. Rather than relying solely on trial and error, organizations could use the framework to design interventions that guide people and AI toward creative ideas.
Authors: Shai Vardi, University of South Florida; Vidyanand Choudhary, University of California, Irvine.