Queensland Artificial Intelligence Risk Index (QAIRI)

What we set out to do: Build the catastrophic risk of AI equivalent to the Australian Disaster Resilience Index (ADRI) at council level, using real data.

What we did: Built six scored components sourced from the International AI Safety Report and Hendrycks' catastrophic risk taxonomy, using real ABS, Queensland government and Queensland Reconstruction Authority data. Used the ADRI model as a comparison to the QAIRI. Grounded the legal case in the actual Disaster Management Act and Human Rights Act and checked against current legislation. Published the QAIRI as an interactive map with adjustable component weights, plus methodology, limitations and legal pages. What we found: QAIRI ranks councils differently to ADRI (r = -0.408). Digitised metro councils rank worse under QAIRI since system dependence is a strength for natural disasters but a liability for AI. One of the six components is genuinely AI-specific and the rest relate to general vulnerability based on the data available for this hackathon.

What's next: Increase number of data sources contributing to the index Gain first hand feedback from local authorities Identify clusters of similar LGAs and promote collaboration, shared resources, action plans Expand to other states and territories