| 摘要: |
| 人工智能技术的广泛应用带来了各种潜在问题，欧盟通过制定《人工智能法案》应对这些问题，但这种“统一立法模式”在应对人工智能时存在难以克服的内在悖论。人工智能的自主性或黑箱特点并未改变法律规范评价的核心焦点，即对行为及其结果属性的关切，现有法律框架足以充分评价和有效控制人工智能的应用及其直接后果，无须额外的统一立法。更重要的是，《人工智能法案》立足于现有欧洲法律价值体系，无法妥当回应科林格里奇悖论式的挑战，因为其忽视了人工智能技术可能引发的深远社会变迁和人类价值重构。应当放弃“一法定乾坤”的立法思路，转而采用更为渐进、灵活的“分散立法模式”，以应对人工智能开发和应用之外更深远的社会后果。 |
| 关键词: 统一立法模式 分散立法模式 科林格里奇悖论 人工智能 新兴技术 |
| DOI： |
| 分类号: |
| 基金项目:国家社会科学基金年度项目“法教义学正当性基础及其实践意义研究”（项目编号： 25BFX002） |
|
| The Future Has Not yet Arrived — The Inherent Paradox of a Unified Legislative Model for AI |
|
Wang Linghao
|
|
|
| Abstract: |
| The widespread application of Artificial Intelligence （AI） technology presents various potential issues. The European Unions AI Act attempts to address these issues through unified legislation, but this “unified legislative model” exhibits inherent and insurmountable paradoxes when dealing with AI. Neither the autonomy nor the black box nature of AI alters the core focus of legal normative assessment — namely, the concern for the attributes of actions and their resulting consequences. Existing legal frameworks are already sufficient to adequately assess and effectively regulate AI applications and their direct outcomes, thereby obviating the need for additional, overarching unified legislation. Moreover, the Act, rooted in the current European legal value system, cannot adequately address challenges akin to the Collingridge dilemma, as it overlooks the profound societal transformations and potential reconstruction of human values that AI technology may induce. The legislative approach of a “one-size-fits-all” law should be abandoned in favor of adopting a more incremental and flexible “decentralized legislative model” to tackle the more far-reaching societal consequences that extend beyond the immediate development and application of AI. |
| Key words: Unified Legislative Model, Decentralized Legislative Model, Collingridge Dilemma, Artificial Intelligence （AI）, Emerging Technologies |
