摘要
生成式人工智能的基础数据来源复杂,数据的真实性与合法性存在极大不确定性,由此出现的虚假信息在技术逻辑上具有必然性和无限复制性;在传播规律上具有瞬时性和隐蔽性;在危害后果上具有复杂性和不可控性。现有多中心协同论、二元治理结构论和层级治理论分别从不同的视角提出治理策略,但过于关注协同合作本身却忽视了治理主体的协同能力和响应效率;过度依赖政府与市场的互补却遮蔽了二者面临新技术时容易懈怠的共性;盲目信任分层治理极易陷入技治主义的困境。对此,应转向敏捷治理下以政府主导强化多元主体间的协同效能、提升治理结构的适应性、破除层级障碍的治理新思路。以虚假信息治理的法治化和技术化为基本方向,保持制度的原则性要求和技术工具价值的协同,实施虚假信息等级评估制度,构建治理主体高效协作机制、快捷响应机制和容错纠错机制,搭建生成式人工智能虚假信息敏捷治理新模式的基本框架。
The basic data sources of generative artificial intelligence are complex and extensive,and the authenticity and legitimacy of the data are greatly uncertain.False information is inevitable and subject to infinite reproduction in the technical logic.It is propagated instantly but it is also concealed and complex and capable of uncontrollable consequences.The multi-center synergy theory,dual governance structure theory and hierarchical governance theory suggest governance strategies from different perspectives.But theyfocus too much on collaborative cooperation and ignore the collaborative abilities and response efficiencies of the governance subjects.They also rely on the complementarity of government and the market to explain their common inclination to become slack in the face of new technologies.A blind trust in hierarchical governance easily leads to the dilemma of technocracy.We should turn instead to a new governance idea under agile governance.We should strengthen collaborative efficiency among multiple subjects under the government’s leadership,improve the adaptability of governance structure through organizational flexibility,and break down hierarchical barriers through penetrating governance.If legalization and technicalization of false information governance are the basic direction,then the principle requirements of the system and the synergy of the value of technical tools are maintained;the false information grade evaluation system is implemented;and,the efficient cooperation mechanism,quick response mechanism,and fault tolerance and error correction mechanism of the governance bodies will be constructed.Ultimately,this will lead to the construction of the basic framework of a new model of agile governance of false information in generating artificial intelligence.
作者
何宇华
李霞
He Yuhua;Li Xia
出处
《治理研究》
CSSCI
北大核心
2024年第4期142-156,160,共16页
Governance Studies
基金
2024年度浙江省哲学社会科学规划课题“公共安全数字化治理的风险识别与法律规制研究”(编号:24NDJC164YB)
浙江省教育科学规划2022年度一般规划课题(高校)“新文科背景下公安网络数据法治人才培养的实践与探索研究”(编号:2022SCG320)
中国社会科学院-上海市人民政府上海研究院交办项目“长三角地区一体化行政执法协作机制研究”(编号:2023jb023)。
关键词
生成式人工智能
虚假信息
敏捷治理
generative artificial intelligence
false information
agile governance
