Public information is a social resource that connects all aspects of social life, regulates social activities and public behaviors, and plays a very important role in influencing public trust. Based on the perspective of communication, we divide the government into two ways to release public information, that is, mass communication and personalized recommendation. Moreover, the public can choose to acquire or not acquire a strategy. Then, this study conducts an evolutionary game between the government and the public to analyze the influence of public information on the public in China. Our objective is to find how the evolutionary path changes between the two players when strategies are changing. Then, a simulation analysis is presented in six scenarios to illustrate the results. The results show that the two players can converge to different stable statuses under different situations.
1 January 2026 in IET Computers & Digital Techniques
The rapid development of generative artificial intelligence (AI) poses new challenges to digital product copyright governance. Considering unauthorized training‐data use, infringing AI‐generated outputs, nonzero infringement costs, and uncertain enforcement outcomes, this paper constructs a two‐population evolutionary game model between generative AI agents and digital copyright holders. The model incorporates infringement costs, probabilistic enforcement, and static and dynamic punishment mechanisms to examine the evolution of compliance, infringement, active enforcement, and passive tolerance strategies. The results show that under static punishment, the system may fall into widespread infringement, periodic oscillations, or passive tolerance. To overcome these limitations, a dynamic punishment mechanism is introduced, in which penalty intensity adjusts according to the AI side’s compliance rate. MATLAB simulations and sensitivity analyses based on normalized but industry‐informed parameters verify the theoretical results. The findings indicate that dynamic punishment can improve system stability and guide periodic fluctuations toward a stable mixed–strategy equilibrium, thereby promoting a more predictable balance between AI compliance and copyright enforcement. This study provides theoretical support for adaptive copyright governance in the era of generative AI.