Guest:
Paul V. DiStefano is a Ph.D. candidate in Cognitive Psychology and Language Science at Penn State University. Originally from Staten Island, NY, he earned his B.S. in Integrative Neuroscience from Binghamton University and his M.S. in Psychology from Penn State. Paul’s research explores how people represent and use information creatively, using behavioral experiments, neuroimaging, and computational methods. He focuses on individual differences in cognition and creativity, as well as developing automated tools for assessing creativity using large language models. His long-term goal is to bridge cognitive science and artificial intelligence to better assess and enhance human creative thinking.
Episode Transcript:
LINDSEY WHISSEL FENTON: Many of us think of AI as a tool that generates content for us, maybe even replacing what we create ourselves. But a new study from Penn State might shift this perception. I'm Lindsey Whissel Fenton with 12 Degrees on WPSU. Today, as we explore the creative degree of wellness, we're going to talk about something that feels especially timely. The relationship between humans, AI and creativity. I'm talking with Paul DiStefano, a researcher at Penn State whose new study suggests that AI can actually help us become better creative thinkers if we use it to stretch our ideas rather than just outsource them. Paul, I'm excited to talk with you about this study, which we should note is currently a preprint. Give us a high-level overview of what the study looked at.
PAUL DISTEFANO: So, our study looked at people generating ideas in an engineering design context, for example reducing air pollution. So, they were coming up with ideas on their own. And then what we had them do was evaluate some AI generated ideas for these problems and then come up with additional ideas. So, we compared people who were engineers and psychology students at Penn State, and what we see is that everyone does better as a result of evaluating AI ideas. But the people who are more creative and had more expertise were the ones who were doing better when we add the AI in.
LINDSEY WHISSEL FENTON: What does that tell us about AI's role in creativity?
PAUL DISTEFANO: This to me tells us that expertise is still important, and creativity is still important. We still need to be focusing on these very human elements. I think a lot of people are getting too worked up on the automation of AI, but one of the biggest takeaways from this entire line of research is that the human element is essential, and we should be focusing on developing systems and frameworks for people to be able to use their own voice. There is a lot of homogeneity in the ideas that people tend to generate when working with AI, so they should be focusing on using AI to improve their voice and their creativity, rather than just accepting what the AI suggests and taking it as fact.
LINDSEY WHISSEL FENTON: If I used AI to prompt feedback for something creative that I was working on, it's not really any different from co-creating with another human.
PAUL DISTEFANO: It can be quite similar to another human. The only thing that is worth noting is that no matter how terrible your idea is, the AI will tell you it's a good idea unless you specifically ask it not to. Another human will tell you when your idea is not that good.
LINDSEY WHISSEL FENTON: How do we make sure to set ourselves up so that we're using AI as a thought partner, without letting it take over the whole creative process?
PAUL DISTEFANO: I think if your goal is to protect your own creative ideation, I think one helpful place to start is by coming up with ideas before you actually work with the model. Generally, if you start and you say, hey ChatGPT, give me some ideas for this project of mine, you're going to get some pretty stereotypical basic answers. So, if you start with your own ideation first and then build on that, you got a lot more creative diversity in the responses that way.
LINDSEY WHISSEL FENTON: What's one thing you'd want the person listening to remember when they're experimenting with AI for creative work?
PAUL DISTEFANO: I'll give you two answers. The first is if any students are using AI to do your homework or to do your creative work, you're not really developing the kinds of skills that make you happy or successful. I mean, I guess this kind of rolls into the second answer I was going to give. Creativity is really enjoyable, actually. People like being creative, so don't give that away to AI. Being creative is fun and you get something out of it.
LINDSEY WHISSEL FENTON: That was Paul DiStefano, a researcher at Penn State whose work explores the relationship between humans, AI, and creativity. This conversation is part of 12 Degrees, a program and podcast from WPSU and the Penn State Nese College of Nursing. Learn more at wellness.psu.edu. I'm Lindsey Whissel Fenton, WPSU.
[END OF TRANSCRIPT]
This is a clip from a longer conversation. To listen to the full episode, visit:
