Over the past decade, communities across the United States have faced a steady rise in severe weather events. Climate Central reports more than 193 major disasters during this period, resulting in over $1.5 trillion in economic losses and 6,403 deaths. Beyond the staggering figures, these events disrupt something even more fundamental: access to basic human rights, including safety, shelter, education and stability.
When disasters strike, the duration and depth of these disruptions are influenced not only by the severity of the event itself but also by how well a community is prepared, how effectively it responds, and how strongly institutions support long-term recovery. As generative AI (GenAI) becomes increasingly embedded in daily life, many are asking an important question: What role can GenAI play in strengthening disaster preparedness and resilience—especially among youth?
Specifically, how might young people ages 12 to 24 be empowered to take an active role in disaster risk reduction using GenAI chatbots? And what educational frameworks would be needed to support such engagement?
A recent study published in Human Rights Education Review by Shuyang Huang, Joshua DeVincenzo and Thomas Chandler, from the National Center for Disaster Preparedness at the Columbia Climate School, addresses these questions. Their work proposes a structured GenAI chatbot framework designed to strengthen youth engagement in disaster risk reduction and identifies the practical features needed to make these tools genuinely empowering.
“GenAI chatbots can help youth engagement in disaster risk reduction education across the phases of preparedness and mitigation, response and recovery. If done effectively, it can provide culturally tailored, proactive education that bridges gaps between school-based learning and community action,” said Huang.
The potential for GenAI chatbots in disaster risk reduction
Real-world examples illustrate how GenAI chatbots are already being used in crisis-related contexts. For instance, Clara (American Red Cross) provides disaster guidance and preparedness information, while ChatClimate offers accessible, interactive explanations of climate risks based on IPCC reports. These emerging tools demonstrate the growing role of chatbots in translating complex information into actionable guidance for the public.
Potential benefits of GenAI chatbots in disaster risk reduction include:
Strengthening youth engagement through bidirectional communication
GenAI chatbots can facilitate two-way information sharing, helping young people not only receive critical updates but also contribute meaningful local data. For instance, UNESCO’s STEDPA project in East Africa enables residents to receive alerts while simultaneously contributing community-generated reports.
Delivering timely, accessible education
GenAI chatbots can quickly synthesize the latest reports and present information in an interactive, age-appropriate manner. Because these chatbots can operate through familiar platforms like SMS and WhatsApp, they can significantly increase the reach and accessibility of essential guidance. However, clear disclaimers and guardrails for AI-generated information are necessary to ensure users understand the limits of automated outputs.
Enhancing community reporting
When young people are encouraged to participate as data contributors, it can improve emergency coordination efforts. Platforms such as Indonesia’s Peta Bencana demonstrate how real-time, youth-generated flood reports can support faster, more accurate responses.
Supporting culturally responsive, inclusive learning
GenAI chatbots can adapt to educational content based on cultural and linguistic needs. With inclusive design, tools can offer multilingual support, age-appropriate messaging and accessibility features, creating equitable learning experiences across diverse ages, genders, languages and abilities. The study emphasizes that such adaptability is especially important for engaging youth with differing communication styles, backgrounds and needs.
While these examples highlight the promise of GenAI chatbots, the study also identifies important limitations in current designs that must be addressed for chatbots to reach their full potential.
Existing gaps and challenges in GenAI chatbots for disaster contexts
As part of the study, the authors conducted a broad review of how current GenAI chatbots perform in disaster-related tasks. This analysis revealed several recurring limitations. Gaps that do not reflect the full potential of GenAI, but rather areas where thoughtful design and youth-centered planning are still needed. These challenges shape how effectively young people can benefit from GenAI during disasters:
Static or limited datasets
Many existing GenAI chatbots rely on training data that may be outdated, incomplete or not locally relevant. When information does not reflect on-the-ground realities, especially in underserved regions, the guidance young people receive may lack precision or context.
User-driven interactions
Most GenAI chatbots still depend on users knowing what to ask and how to phrase it. Young people may struggle to pose clear or specific questions, which can limit the usefulness of chatbot responses without structured prompts or guided interaction.
Risk of misinformation
GenAI chatbots can occasionally generate incorrect or misleading content, particularly in fast-moving disaster situations. For youth who may lack the expertise to critically evaluate responses, this risk is heightened.
Persistent accessibility barriers
Reliable internet access, stable electricity, and linguistic inclusivity remain major obstacles. Many GenAI chatbots do not yet support minority languages or dialects, which can reinforce existing inequities.
Collectively, these challenges point to the need for a more structured, youth-centered approach to GenAI design in disaster settings.
What a youth-centric GenAI chatbot framework should prioritize
Building on the gaps identified in current GenAI chatbots, the study proposes a set of priorities that would guide the design of a more reliable, accessible, and youth-centered chatbot. Rather than treating these challenges as limitations of GenAI itself, the framework outlines practical design elements that can make chatbots more supportive of young people across preparedness, response and recovery.
To ensure GenAI strengthens, rather than compromises, disaster resilience and youth empowerment, the authors propose several enhancements. The items below represent a selected set of priorities drawn from the study’s larger, more detailed framework:
- Real-time localization and personalized guidance
- Multi-modal information delivery (e.g., maps, videos, graphics, audio)
- Two-way reporting mechanisms
- Integration with official emergency systems and verified databases
- Psychological first aid and emotional support, including crisis-detection capabilities and referrals to emergency hotlines
- Notifications that support long-term civic participation
- Tools that promote ongoing engagement and recovery monitoring
- Youth-friendly features such as simple prompts, age-sensitive messaging, and opportunities for young people to contribute information back to the chatbot.
Together, these features illustrate what a youth-centric GenAI chatbot should prioritize to effectively support decision-making, emotional well-being and community engagement during disasters.
However, the authors emphasize that strong ethical safeguards must accompany these advancements. GenAI should not replace human-support systems, nor should it inadvertently cause harm through surveillance, bias or misinformation. Emotional well-being and user privacy are critical considerations, especially when supporting young people during and after crises.
“GenAI chatbots, if thoughtfully designed and equitably implemented, can serve as a transformative tool for strengthening youth participation in disaster risk reduction,” said DeVincenzo.
If developed intentionally, a GenAI chatbot’s ability to scale thousands of users, operate continuously and provide culturally responsive guidance can strengthen both immediate disaster response and long-term community resilience.
“By embedding GenAI chatbots into education, emergency communication and long-term recovery systems, we can build more inclusive, resilient applications that will increasingly add value to communities in need,” Chandler said.
Click here to read this full study.
The National Center for Disaster Preparedness (NCDP) works to understand and improve the capacity to prepare for, respond to, and recover from disasters. NCDP focuses on the readiness of governmental and non-governmental systems, the complexities of population recovery, the power of community engagement, and the risks of human vulnerability.
Shuyang Huang is a staff associate II working on data management and analysis at NCDP.
Josh DeVincenzo is the assistant director for education and training and adjunct assistant professor at the NCDP.
Thomas Chandler is the managing director and research scientist at the NCDP and an associate member of the Columbia Climate School faculty.
