A combined approach of evolutionary game and system dynamics for user privacy protection in human intelligence interaction

The rapid development of generative artificial intelligence (GenAI) has generated significant economic and social value, alongside risks to user privacy. For this purpose, this study investigates privacy protection in human-AI interaction by employing a combined approach of evolutionary game and system dynamics. A three-party game model was developed to analyze the interactive effects and evolution of privacy protection strategies among the government, GenAI company, and users. Sensitivity analysis through system dynamics simulations was conducted on four kinds of factors-government, company, users, and incentive mechanisms, to reveal how these factors influence the strategy choices of the three parties. The results suggest that the government's reputation, subsidies, free-riding benefits, fines, rewards from GenAI company to users, and the cost-benefit considerations of all three parties are key factors affecting strategic decisions. Moderate fine and subsidy policies can effectively promote privacy protection, with subsidy policies proving to be more effective than penalty policies. This paper provides theoretical support and decision-making guidance for balancing technological development and privacy protection in human-AI interaction, contributing to the regulated and orderly development of Generative Artificial Intelligence.