重庆理工大学学报2025，Vol.39Issue(18)：91-102,12.DOI:10.3969/j.issn.1674-8425(s).2025.09.009
AI安全国际监管合作的困境与解决构想
Generative AI security:Challenges and pathways for international collaborative governance
摘要
Abstract
As generative AI technology achieves breakthroughs in capability continiously and opens its APIs fully,generative AI has enhanced content generation capabilities while simultaneously proliferating inherent risks of black-box opacity,uncertainty,and ubiquity,extending artificial intelligence security concerns from technological ethics to national and global strategic dimensions.However,current international AI safety regulation demonstrates insufficient effectiveness in addressing these risks,attributable to three primary factors.First,at the normative level,ethical initiatives and soft law mechanisms constitute the primary governance tools,yet their non-binding nature fails to establish effective constraints;secondly,paradigmatic conflicts in regulatory philosophies among sovereign states exacerbate normative fragmentation trends;thirdly,deep involvement of non-state actors creates complex strategic interactions.Given the increasingly complex and dynamic challenges facing AI governance,singular governance models seem inadequate for effectively addressing multifaceted risks.Therefore,there is an urgent need to develop the concept of artificial intelligence sovereignty while respecting the principle of state sovereignty,establish an umbrella governance structure coordinating hard and soft law mechanisms,and construct cross-border data governance frameworks leveraging international digital economy rules,thereby building a fair,transparent,inclusive,and orderly new paradigm for international AI governance.关键词
生成式人工智能/人工智能安全/国际监管合作困境/软硬法协同治理Key words
GenAI/AI security/international regulation challenge/coordinated governance combining hard and soft law分类
社会科学引用本文复制引用
韩龙,程千帆..AI安全国际监管合作的困境与解决构想[J].重庆理工大学学报,2025,39(18):91-102,12.基金项目
国家社会科学基金重大项目"数字化背景下国际金融服务的法律治理体系研究"(22&ZD203) （22&ZD203）
