Published: Nov. 3, 2023 By

Stepping into the Center for the Brain, A.I., and Child (BAIC), discussions vary from how ChatGPT is influencing parenting, to using generative AI art for student expression to how children could use chatbots for storytelling and what culturally sensitive AI development means. 

The lab's eclectic, cross-disciplinary environment focuses on how AI will impact the next generation of children and their caretakers around the world, as generative AI becomes a new normal. 

The lab has two centers. One, at the University of Colorado Boulder, is headed by Tom Yeh, a professor of computer science and the other center is led in close collaboration by Pilyoung Kim, a professor of psychology at the University of Denver.  

Hopes for the research

The hope of the lab, Yeh explained, is that by asking and answering as many questions as they can around child development and AI, the lab can encourage ethical guardrails and culturally-specific AI protocols. 

"Beyond ChatGPT, I'd like us to think about how our children are developing in this environment," Yeh said. "It used to be that children didn't have to watch for cars when they crossed the street, but suddenly they had to learn, and it's the same with AI. There's a reconfiguration of our society happening," he explained. 

Let's meet some lab members to get a better sense of what the BAIC lab is working on. 

Cornelius Adejoro, 2nd Year PhD student

Cornelius Adejoro (PhD'27) is a second year PhD student whose work focuses on culturally sensitive AI and technology for vulnerable children and those in the global south

Explaining what culturally sensitive AI means for him, he said, "It means inclusion, it means fairness, incorporation and an equitable society–helping others to make maximum use of these technologies by upholding their cultural values." 

One of Cornelius Adejoro's first projects in the lab was an essay contest with students in Nigeria, where he engaged middle school students in Nigeria through a writing contest to share their thoughts on what they hope AI can do for them as a way of shaping the future of AI. He explained that many students were hopeful for better translation between languages. 

"Nigeria has over 500 spoken languages, so a class of 50 students could have 50 languages," he said. 

Cornelius Adejoro said that being part of the lab has been a positive experience. "Tom creates a system that keeps you going. We meet often and I have a receptive advisor who is ready to listen, whatever the idea is. Even if we have individual work, we often meet in teams to allow for collaboration," he said. 

Larissa Schwartz, 1st year PhD student

Larissa Schwartz is both part of BAIC and the ATLAS Institute. Schwartz is a professional graphic designer, and was previously a visual arts and design instructor for public education in Denver and Houston. She said she began researching and wondered if generative AI art was being used ethically in education. 

She was able to work with the BAIC lab to pilot a generative AI art competition for more than 50 high school students in Aurora Public Schools. The students created art through traditional mediums and then used that same artistic idea to prompt DALLE-2, an AI system that can create images and art from a written description. The two pieces were then displayed next to each other and the students’ work was presented at the ATLAS Institute's projects expo in May.

Last year, Schwartz completed her master’s degree from the ATLAS Institute, with Yeh as her thesis advisor before she began her PhD this year. 

"He connected me to people who were so helpful with the things I wanted to build and create, and he really listened to the things I wanted to do as a student and in the future, in terms of my PhD and working in industry," she said. 

Seungwook Lee, 2nd year Masters student 

Seungwook Lee is part of a research project in BAIC that questions what the right way to use Generative AI in future generations is. 

"We want to minimize the potential harm to children's education," he said, explaining that data suggests over 40% of teachers in the United States are already using ChatGPT in their classrooms

The project is to use generative AI to bring out children's creativity by using storytelling. He is working on the part of the research that controls generative AI prompts to be emotionally safe and creative for children.

"It's important to know first if using these models impacts the child, before deciding how to go ahead," he said. 

The goal is to study how large language models affect children using brain scans and interviews. Based on the results, Lee will be designing methods for fine-tuning a large language model to more safely support children's creativity.