From Blade Runners to Tin Kickers: what the governance of artificial intelligence safety needs to learn from air crash investigators

What should we do when artificial intelligence (AI) goes wrong? AI has huge potential to improve the safety of societally critical systems, such as healthcare and transport, but it also introduce new risks amplify existing ones. For instance, biases in widely deployed diagnostic systems could adversely affect care a large number patients (Fraser et al. 2018), hidden weaknesses perception autonomous vehicles may regularly expose road users significant risk (NTSB 2019). are most appropriate strategies for governing AI-based systems? One answer emerges from taking contrasting looks forwards our imagined dystopian future backwards progressive evolution aviation safety.