Abstract: As artificial intelligence becomes increasingly ingrained in the realm of public security and governance, its practical application transcends unidimensional technological innovation. It now encompasses the development of a multi-stakeholder collaborative governance model, compels the upgrading of supporting public security systems, and drives the transition of governance methods towards more technologically advanced means. Simultaneously, the rapid advancement of intelligent technologies has resulted in structural imbalances among governance actors, the inadequacy of governance systems, and technological alienation, giving rise to risks such as social inequality, disarray within governance systems, and an overreliance on technology. Therefore, it is crucial to elucidate the evolving interplay among various actors in the context of artificial intelligence embedded in public security governance, while strengthening forward-looking prevention, constraint guidance, and formulating effective risk mitigation strategies. This approach is of both theoretical and practical value for harnessing artificial intelligence to enhance public security management more effectively.
This article establishes an evolutionary game model involving three key stakeholders: Government departments (represented by public security agencies), artificial intelligence companies, and the general public. And then it analyzes behavior strategies of the tripartite stakeholders in the scenario of artificial intelligence embedded in social security governance. By examining the strategic combinations chosen by these three interest groups, a game payoff matrix is constructed. Through the analysis of three-party replicator dynamic equations, the equilibrium points and stability of the evolutionary game among the three parties are determined.
The findings indicate that government departments, exemplified by public security agencies, must proactively coordinate the structural relationships among various stakeholders in the realm of artificial intelligence integration. It is crucial to foster a collaborative governance framework involving multiple entities, enhance and refine supporting systems, and expedite the development of an AI technology governance structure centered on “social good governance”. This approach will comprehensively facilitate the positive progression of AI integration within social security management.In particular, government departments represented by public security agencies need to take the initiative to comprehensively coordinate the structural relationship among multiple subjects under the embedding of artificial intelligence, promote the formation of a multi-subject collaborative governance pattern, follow up and improve supporting systems, and accelerate the construction of artificial intelligence for “good social governance”. The intelligent technology governance system can comprehensively promote the benign development of artificial intelligence technology embedded in social security governance.
This paper delves into the optimal risk response strategy within the evolutionary game of the interests of three key stakeholders, beginning with the logic of security management and adopting a more comprehensive and systematic perspective. Nevertheless, there exist certain research limitations. Future studies will further deepen the focus in two areas: First, incorporating relevant regulatory authorities to establish a four-party evolutionary game model, in order to develop more effective multi-stakeholder governance strategies; Second, investigating the creation of a more practical game payoff matrix, offering superior solutions for boosting the technical security and stakeholder collaboration of AI integration in social security management.

Key words: Artificial Intelligence, public security governance, risk analysis, collaborative co-governance, social governance, tripartite game

摘要： 随着人工智能在社会治安治理领域应用日趋深入,其嵌入治安治理的实践逻辑已不再是单向度的技术创新,而是包括推动形成多元主体协同共治模式、倒逼治安治理配套制度体系升级、驱动治理手段技术化转型的多向度赋权。与此同时智能技术加速升级所引发的治理主体结构性失调、治理制度失效,以及技术异化,也产生了社会公平被破坏、治理体系混乱、过度技术依赖等风险。为寻得风险因应之道,本文建立以公安机关为代表的政府部门、人工智能企业、公众主体三方利益相关主体的演化博弈模型,并对模型演化策略进行分析。结果表明:以公安机关为代表的政府部门需要主动作为,全面统筹人工智能嵌入下多元主体间的结构关系,推动形成多元主体协同共治格局、跟进完善配套制度、加快构建“社会善治”的人工智能技术治理体系,以全方位促进人工智能嵌入社会治安治理良性发展的因应之道。

关键词: 人工智能, 治安治理, 风险分析, 协同共治, 社会善治, 三方博弈