Emergency managers consider using AI chatbots to draft disaster alerts

Emergency managers are considering using AI chatbots to help draft public alerts during disasters, since humans tend to communicate poorly about emergencies by using jargon, leaving out needed information, or including distracting details. While most government agencies currently don't permit employees to use generative AI, the technology is already spreading through state and local disaster departments, with hundreds of companies marketing AI tools for emergency management tasks. Poor emergency communication can have deadly consequences, as shown by past cases where unclear or contradictory warnings led to confusion and loss of life during wildfires and mudslides. The challenge is that while AI systems could theoretically help by rapidly synthesizing best practices from thousands of past alerts, emergency managers and experts remain uncertain about whether the public will trust AI-generated warnings and whether the technology is accurate enough to deploy in real emergencies.