A digital technology-empowered self driving car seems to be a fascinating invention that is designed and to evolve through to some extent collective human senses, which is deep learning. 
Yet the reliability still remains in doubt as operation of the digital artefact is still counted within human-made systems and senses, which is human-made algorithm. It may perhaps account for predictable risks that can be addressed within accumulated existing data that has been dealt within human-made worlds. 
However, unlike the confident speech made by presenters working at one of the biggest commercial companies in the world, what if the digital artefact, self-driving car are faced completely 'uncertain' situations beyond considerations of those human- made algorithm ; and what if the advanced software system cannot be perfectly supported by its physical partner which is hardware when it is confronted with such un-expectable situations?