As artificial intelligence continues to grow and seeps into spaces like art, design and writing, a Houston researcher is examining its effects on creativity.
University of Houston’s Bauer College Assistant Professor Jinghui Hou, in collaboration with scholars around the world, recently published the paper "The Double-Edged Roles of Generative AI in the Creative Process" in the journal Information Systems Research.
Through the research, the team identified two stages of creativity that AI can influence: ideation and implementation.
In one study, Hou and her team developed a lab experiment to examine the impact of a cutting-edge generative AI tool during the brainstorming or ideation phase on a group of designers with varying levels of expertise.
The study showed that nearly all designers who used generative AI during this stage improved in the creativity of their graphic design work, and that the improvements were substantial and consistent across the board.
“In the first stage, we find that for anyone, including ordinary people and expert designers, AI is very helpful because of its computational power,” Hou said in a news release. “It can go beyond the imagination that humans have. For example, if I wanted to imagine a tiger with wings, it would be hard to see that in my head, but AI can do it easily.”
However, a second study examining the implementation stage found that AI affects professionals differently than novice designers.
The study showed that novice designers continued to improve in all aspects of their work when using AI. But more expert designers did not see significant improvements in the implementation stage. Rather, expert designers who used AI spent 57 percent more time completing their work compared with their peers who did not use AI.
“In the implementation stage, we find that AI is still very helpful for those ordinary people, but it creates more work for expert designers,” Hou said in the release. “This is because the designer has years of training to materialize a piece of artwork. We find that AI uses different techniques to produce creative work. For designers, it can become burdensome to revise what AI made.”
Hou’s paper suggests that AI is most helpful in the brainstorming stage, but hopes to see generative AI developers program tailor the technology for expert-level, professional needs.
“It could give users more freedom to fit the technology to their usage pattern and workflow,” Hou added. “In a sense, it's not about people catering to the AI, but the AI technology catering to people."