AI trained for weather warnings
An AI chatbot could improve disaster response with personalised emergency information.
Developed by postgraduate students at James Cook University, the chatbot provides postcode-specific disaster updates, tailored to different locations and types of emergencies.
The concept won the 2024 Disaster Challenge, hosted by the Department of Fire and Emergency Services Bushfire Centre of Excellence.
Team member Madison Green said their personal experiences with disasters, such as Cyclones Yasi and Kirrily, and the 2019 Townsville floods, drove their idea.
“We experienced a lack of personalised messaging around the risks, which left us feeling uncertain and unsure about how to help ourselves and others,” Green said.
This inspired their focus on customised information to build resilience and trust in communities.
Alison Sheaves, another team member, noted the need for more relevant disaster information.
“An AI chatbot providing tailored, relevant messaging would fill this gap,” she said.
The chatbot could help people with preparation, offer real-time updates, and give advice on risk mitigation during disasters.
The students are looking to develop the idea further with industry collaboration.
Natural Hazards Research Australia CEO Andrew Gissing says the project shows how non-traditional disaster research can build trust between communities and emergency services.
“Harnessing technological advancements like AI could provide certainty during uncertain times, building trust and enhancing resilience,” he said.
The Disaster Challenge encourages early-career researchers to tackle disaster management issues, with other finalists pitching ideas on heatwave engagement and home insurance incentives.