Im surprised we dont hear more about AI fails already - the odd db getting dropped but not much else
Plenty of stuff out there already about end users trusting it too much and leading to loss (of money, health, life, etc) and bias affecting minority groups (mostly from researchers and the public sector).