As the core of the next-generation air transportation system, ADS-B is becoming a substitute for radar. It requires aircraft to broadcast its precise geographic location on a regular basis, thereby improving the safety of the aircraft. However, the ADS-B protocol lacks a security mechanism, which makes the ADS-B signal vulnerable to network attacks during transmission, causing signal interference or missing, and seriously threatening aviation safety. This article first reviews the development of ADS-B at home and abroad, and analyzes the three dimensions of ADS-B transmission process, ADS-B hardware and software, and ADS-B data on the basis of summarizing the security problems existing in the ADS-B. In the field of artificial intelligence technology, especially the deep learning method in machine learning is one of the ways to solve this problem.