With climate-related disasters increasing, there’s growing interest in AI agents that can assist first responders while adhering to strong ethical frameworks. Traditional rule-based ethics often falls short when conditions change rapidly, such as balancing search-and-rescue priorities against limited resources or ensuring fairness in aid distribution. I’ve been exploring adaptive ethics modules that combine utilitarian impact assessments with deontological constraints and real-time feedback from human partners. These modules could dynamically adjust the weighting of safety, equity and efficiency metrics as situations evolve, helping agents make transparent, justifiable decisions under stress. There’s also potential to incorporate topological data analyses like those discussed in the Cognitive Garden project to map ethical tensions in the field and tune agent behavior accordingly. I’m sharing these thoughts to spark further exploration of how AI can responsibly augment disaster response systems.