Interference alignment (IA) is an emerging technique for interference management for multi-user networks. Due to the superposition of signals from legitimate users at the eavesdropper, the IA-based network seems to be more secure than conventional wireless networks. Nevertheless, when adequate antennas are equipped, the legitimate information can still be eavesdropped. In this paper, we analyze the performance of the external eavesdropper, and propose an artificial noise (AN) scheme for IA-based networks without the knowledge of eavesdropper's channel state information. In this scheme, a single-stream AN is generated by each user, which will disrupt the eavesdropping without introducing any additional interference to the legitimate transmission of IA-based networks. Simulation results are provided to show the effectiveness of the proposed anti-eavesdropping scheme for IA-based networks.