Bouncing back or forward?: artificial intelligence in disaster resilience practices

Scholars across many disciplines have turned their focus toward resilience, or the capacity to bounce back after a disruption. Resilience has been promised as a positive process for recovery from climate-related disasters. Less discussion has focused on the ‘dark side’ of resilience, or the capacity to bounce back to pre-disruption patterns that are destructive. This paper considers recent climate disasters and corresponding resilience efforts, tracing the infrastructure of recovery to ask how negative patterns of resilience may form. In particular, new applications of artificial intelligence may promise more efficient disaster recovery while perpetuating inequities. The paper engages a case study of the Federal Emergency Management Agency (FEMA)’s use of AI to conduct disaster damage assessment as well as the pitfalls of AI in disaster recovery. AI has generative potential in resilience efforts, but can also reinforce resilience thinking that works against, rather than with, nature.