Great point! The development of AI is always controversial. On the one hand, it helps our lives easier and more intelligent. On the other hand, it also threats our lives as we may lose certain level of control from it. The key issue is that we don't allow AI to make mistakes, especially deadly mistake. For the self-driving cars, although it actually reduces the risks of traffic accident, but once any accident happens, we will be mad at it, regardless of it's a random issue or system fraud. For the weapon control, if AI helps us make decisions, from the statistically point, it will calculate and make the most "accurate" decisions, but any negative results will change our views toward it. However, any AI system is built based on the data and algorithm we human beings input. Therefore, it is still our mindset, how we want to make this world run that decide when the AI should go.