探索与争鸣 ›› 2024, Vol. 1 ›› Issue (8): 105-118.

• 技术与文明 • 上一篇    下一篇

新兴人工智能的政治影响与治理策略

王中原
  

  • 出版日期:2024-08-20 发布日期:2024-09-18
  • 作者简介:复旦大学社会科学高等研究院副教授、国家发展与智能治理综合实验室(教育部)兼职研究员。(上海 200433)
  • 基金资助:
    国家社科基金青年项目“算法治理视域下政府数字化转型的实践困境及其破解机制研究”(21CZZ039)

The Political Impact and Governance Strategies of Newly Emerging Artificial Intelligence Technologies

Wang Zhongyuan
  

  • Online:2024-08-20 Published:2024-09-18

摘要:

新兴人工智能将深刻重塑当代政治,催生一系列前沿的政治学研究议题。新兴人工智能通过重构信息机制、交互机制、决策机制,对政治系统产生深远影响,突出表现在政治输入(选举政治、协商政治)、政治过程(政治代表过程、公共政策过程)、政治输出(政治回应、政治问责)三个环节及其六大场景中。人工智能技术并非完全中性,新兴人工智能的内在技术属性决定了其内嵌的功能和风险,这些功能或风险会在不同的人机关系结构中被放大或遮蔽,呈现出多样的组态和生态。政治场景相较商业场景更具敏感性和冲突性,技术风险极易扩散为政治风险,对国家统治、政民关系和社会稳定产生冲击。然而,从“智能技术”到特定“政治后果”并非因果必然,其间受到制度环境、应用场景和治理策略等多重因素的影响。规范人工智能的政治应用可从开发者、使用者、监管者三方着手,并结合“技术路径”与“关系路径”展开。深度智能呼唤深度民主。

关键词:

Abstract:

Newly emerging artificial intelligence (AI) will significantly shape modern politics and give rise to a series of cutting-edge research topics in political science. The new generation of intelligent technologies, by reshaping information mechanisms, interaction mechanisms, and decision-making mechanisms, profoundly influence political systems. This influence is particularly evident in three stages and six major scenarios: political input (electoral politics, deliberative politics), political process (political representation process, public policy process), and political output (political responsiveness, political accountability). AI technologies are not entirely neutral. The inherent technical attributes of newly-emerging AI determine its embedded functions and risks, and these functions or risks will be amplified or obscured in different human-machine relationship structures, presenting diverse configurations and digital ecologies. The political arena is more sensitive and conflictual compared to the commercial domain, where technological risks can easily spill over into political risks, impacting state governance, state-society relations, and social stability. However, the path from “intelligent technology” to specific “political consequences” is not causally inevitable. It is influenced by multiple factors, including institutional conditions, application scenarios, and governance strategies. Regulating the political use of AI can be approached from the perspectives of developers, users, and regulators, and should be pursued through both the “technological path” and the “relational path”. Deep intelligence calls for deep democracy.

Key words: