NSF awards researchers $1.3 million grant to strengthen Latino children’s STEM learning through interactive, AI-generated e-books
By Carol Jean Tomoguchi-Perez
August 15, 2024
August 15, 2024
Young children are primed to learn science best when they actively engage with meaningful topics that connect to their everyday lives. Focusing on culturally relevant storytelling and parental codesign to increase equity within the Latino community, the National Science Foundation (NSF) awarded professors Mark Warschauer (PI) and Andres Bustamante (co-PI) $1.3 million for their initiative “Collaborative Research: Advancing Latino Children’s Science Learning through Community Co-Design of AI-Enhanced Bilingual E-Books.”
Artificial intelligence can be helpful for developing science learning content quickly and making it interactive; at the same time, there is a risk of inherent racial, social and linguistic biases in AI-generated materials. The four-year project aims to mitigate that risk by directly involving parents in codesigning 24 Spanish-English, culturally relevant e-books for Latino children from four- to seven years old. Parental involvement in the project can help center the content of the e-books in families’ cultural assets and reduce threats of stereotype and bias in the material. This co-design activity will be carried out with families from the Santa Ana Early Learning Initiative, a long-term partner of Warschauer and Bustamante. The initiative utilizes storytelling, a form of cultural capital in Latino communities, to engage children’s scientific curiosity at a young age, when the development of skills and knowledge for success in STEM is critical. The project also aims to build the community’s AI literacy skills. It will use participatory design with parents in underresourced Latino families in California and Michigan to create the e-books, which will feature an AI-powered bilingual conversational agent that enables children to interact with the stories. The e-books will also include family discussion prompts to encourage parent-child interaction. |
Once the e-books are piloted and iteratively improved, a randomized control trial will be carried out with 120 Latino families to evaluate the impact of the e-books on children's science knowledge and engagement, and on parent-child science communication. After evaluation and further improvements, the e-books will be nationally distributed at no cost, significantly broadening STEM learning opportunities for diverse children, particularly in Latino communities.
The total funding for the project is $2 million, which includes a collaborative award to the University of Michigan. That portion is led by former UC Irvine Ph.D. in Education student Ying Xu, who served as an assistant professor at Michigan before recently taking a faculty position at Harvard University.
The total funding for the project is $2 million, which includes a collaborative award to the University of Michigan. That portion is led by former UC Irvine Ph.D. in Education student Ying Xu, who served as an assistant professor at Michigan before recently taking a faculty position at Harvard University.
Abstract:
Early childhood years are critical for developing the foundational knowledge, skills, and attitudes for later success in STEM. Young children learn science best when they actively engage with topics that are meaningful to their everyday lives. Artificial intelligence can help in developing science learning content and making it more interactive, but inherent social, racial, and linguistic biases in AI-generated materials make this undertaking risky. Our project involves direct participation from parents in under-resourced Latino families in co-designing AI-based educational materials. Their participation can help reduce potential biases in the produced materials. University and community partners will jointly work with AI to first create bilingual science stories rooted in Latino identity and then to make those stories interactive. This initiative leverages storytelling–a major form of cultural capital in the Latino community–to foster children's scientific curiosity and engagement, while also helping build community members’s AI literacy skills. The project will contribute important knowledge about how AI can be effectively and equitably harnessed by and with diverse communities in support of their values and education, aligning with key National Science Foundation objectives.
The project utilizes participatory design with Latino families in California and Michigan to create 24 Spanish-English culturally relevant e-books for Latino children aged 4-7, employing generative AI for rapid, iterative content development. The e-books will feature a bilingual AI-powered conversational agent that allows children to dialogue directly with the story characters, as well as family discussion prompts to encourage parent-child interaction. After the 24 interactive e-books are piloted and iteratively improved, a randomized control trial will be carried out with 120 Latino families to evaluate the impact of e-book use on children's science knowledge and engagement and on parent-child science communication. Subsequent improvements will prepare the e-books for free national distribution, significantly broadening STEM learning opportunities for diverse children, particularly in Latino communities.
This Integrating Research and Practice Project is funded by the Advancing Informal STEM Learning (AISL) program, which supports research on the development and impact of STEM learning opportunities in informal educational environments.
Early childhood years are critical for developing the foundational knowledge, skills, and attitudes for later success in STEM. Young children learn science best when they actively engage with topics that are meaningful to their everyday lives. Artificial intelligence can help in developing science learning content and making it more interactive, but inherent social, racial, and linguistic biases in AI-generated materials make this undertaking risky. Our project involves direct participation from parents in under-resourced Latino families in co-designing AI-based educational materials. Their participation can help reduce potential biases in the produced materials. University and community partners will jointly work with AI to first create bilingual science stories rooted in Latino identity and then to make those stories interactive. This initiative leverages storytelling–a major form of cultural capital in the Latino community–to foster children's scientific curiosity and engagement, while also helping build community members’s AI literacy skills. The project will contribute important knowledge about how AI can be effectively and equitably harnessed by and with diverse communities in support of their values and education, aligning with key National Science Foundation objectives.
The project utilizes participatory design with Latino families in California and Michigan to create 24 Spanish-English culturally relevant e-books for Latino children aged 4-7, employing generative AI for rapid, iterative content development. The e-books will feature a bilingual AI-powered conversational agent that allows children to dialogue directly with the story characters, as well as family discussion prompts to encourage parent-child interaction. After the 24 interactive e-books are piloted and iteratively improved, a randomized control trial will be carried out with 120 Latino families to evaluate the impact of e-book use on children's science knowledge and engagement and on parent-child science communication. Subsequent improvements will prepare the e-books for free national distribution, significantly broadening STEM learning opportunities for diverse children, particularly in Latino communities.
This Integrating Research and Practice Project is funded by the Advancing Informal STEM Learning (AISL) program, which supports research on the development and impact of STEM learning opportunities in informal educational environments.