Interdisciplinary faculty team advances human‑centered AI to empower tribal nations in flood planning
A cross‑disciplinary team at the University of Texas at Austin is researching how AI can better support American Indian and Alaska Native nations in flood preparedness. Assistant Professor Lidia Cano Pecharromán in the School of Architecture and Assistant Professor ChangHoon Hahn in the Department of Astronomy are redesigning an AI tool for flood evaluation so tribal nations can incorporate Indigenous worldviews and community‑focused measures of effectiveness into their mitigation programs.
This project will be the first of its kind to re-examine an AI tool from multiple lenses to serve indigenous world views. It’s critical research at a time when flood hazards are intensifying for Native American communities.
The project is motivated by past research that has shown that Native American lands face increasing flooding risks due to the frequency of extreme precipitation events, the decommissioning of dams, and the flood-prone areas they inhabit due to their historical displacement to less desirable geographical areas. In response to flood threats, the federal government created programs to promote flood management best practices across the US; however, it is estimated that only 7% of American Indian and Alaska Native nations participate in these flood preparedness initiatives. The lack of evidence of whether these programs work and can serve the needs of these nations constitutes a barrier for Native American councils to act.
“What we don’t measure in our models is rendered invisible and can lead to inappropriate decision-making,” said Lidia Cano Pecharromán, assistant professor in the School of Architecture.
AI architectures can serve as valuable tools to make up a more holistic view of flood management. Cano-Pecharromán and Hahn are developing a tool to appraise the effectiveness of flood mitigation programs from a relational lens, able to reflect the priorities and cosmovision of American Indian and Alaska Native nations.
“AI tools can be very technically accurate but at the same time unfit to capture a particular world view or objective," said ChangHoon Hahn, assistant professor in the Department of Astronomy, College of Natural Sciences.
Cano-Pecharromán and Hahn’s project is one of six awarded seed funding through Good Systems’ internal competition, aimed at advancing interdisciplinary research in human-centered AI. With the seed funding from Good Systems, the team will expand their research and organize a workshop to bring different scholars and experts to discuss the findings at UT Austin later in the fall.
“Good Systems is honored to support this stakeholder-centered research to advance flood mitigation and better serve our communities,” said Dr. Kenneth R. Fleischmann, Interim Dean of the School of Information and Chair of Good Systems. “We are excited to welcome Drs. Cano-Pecharromán and Hahn and their research team to our network, which brings together leaders across disciplines and sectors to define, evaluate, and build AI systems that benefit society.”
This research project is just taking shape, with additional developments forthcoming. Together with colleagues in a partner university and with tribal councils from different parts of the country, Cano-Pecharromán and Hahn are also seeking funding for a more ambitious project with two purposes:
Re-examine existing AI tools that are being employed by American Indian and Alaska Native nations for forest management and climate adaptation from a holistic lens;
Examine and propose the necessary changes in Federal Indian Law to anticipate legal requirements for any AI tool required by the federal government to embed such holistic worldviews as applied to Native American lands.
This unprecedented effort will be key to guaranteeing responsible use of AI tools that can serve multiple ways of knowing and understanding the human-nature relationship and that uphold the needs of American Indian and Alaska Native nations.