首页|期刊导航|阅江学刊|超级治理的超级隐忧

超级治理的超级隐忧OA

The Super Concerns of Super Governance

中文摘要英文摘要

超级智能是各个方面能力全面超越人类的超级AI,整个社会完全在超级AI治理下运转即超级治理.超级治理依托的社会模拟器技术在思路上存在根本缺陷,它无法区分社会实然与应然,人类行为及其约束条件也无法完全函数化模拟.人身心的双重机器化问题,也将在超级治理下彻底恶化.但超级智能概念仍然能作为AI觉醒类术语和人文批判法的"理想型",发挥其反面启示法的作用.对超级治理必须实施再治理,其关键在于监管AI研发者.

Superintelligence refers to an AI that comprehensively surpasses humans in all aspects.Super governance occurs when society operates entirely under the governance of such super AI.The social simulator technology underpinning super governance suffers from a fundamental conceptual flaw.It cannot distinguish between the actual and the ideal state of society,nor can it fully model human behavior and its constraints as functions.The dual mechanization of human body and mind will also deteriorate drastically under super governance.Nevertheless,the concept of superintelligence can still serve as an"ideal type"within AI awakening discourse and humanistic critique,functioning as a negative example.Super governance must be subjected to Re-governance,with the key lying in regulating AI developers.

刘永谋;王亮家

中国人民大学哲学院||中国人民大学人工智能治理研究院中国人民大学哲学院

超级智能超级治理社会模拟器人的机器化

《阅江学刊》 2026 (1)

73-79,7

国家社会科学基金重大项目"现代技术治理理论问题研究"(21&ZD064).

评论