In this paper, we propose a cooperative approach to improve the security of both primary and secondary systems in cognitive radio multicast communications. During their access to the frequency spectrum licensed to the primary users, the secondary unlicensed users assist the primary system in fortifying security by sending a jamming noise to the eavesdroppers, while simultaneously protect themselves from eavesdropping. The main objective of this work is to maximize the secrecy rate of the secondary system, while adhering to all individual primary users' secrecy rate constraints. In the case of active eavesdroppers and perfect channel state information (CSI) at the transceivers, the utility function of interest is nonconcave and the involved constraints are nonconvex, and thus, the optimal solutions are troublesome. To solve this problem, we propose an iterative algorithm to arrive at least to a local optimum of the original nonconvex problem. This algorithm is guaranteed to achieve a Karush-Kuhn-Tucker solution. Then, we extend the optimization approach to the case of passive eavesdroppers and imperfect CSI knowledge at the transceivers, where the constraints are transformed into a linear matrix inequality and convex constraints, in order to facilitate the optimal solution.