The advent of a new technology is inevitably accompanied by new security and safety issues. As mentioned in Chap. 3 , due to immaturity and inherent defects, a new information technology usually gives rise to two new types of security and safety concerns. One is with the system itself, caused by the inherent vulnerability of the new technology, which is coined the “AI security” problem. The other refers to the resultant safety issues in other areas, although the vulnerability of the new technology does not cause problems to the system itself. And we name this as the “AI safety” problem. In terms of AI security, it partly originates from some loopholes in the new technology, and most of the time these can be detected and eliminated; in another occasion, the new technology may have a natural defect, which makes the problems inevitable and unsolvable by improvement, and the problems can be only circumvented by taking other measures.