首页|算法产业促进与治理法(建议稿)

算法产业促进与治理法(建议稿)

扫码查看
算法是人工智能乃至智能化的核心技术,与算力、数据一起被称为智能时代或数字经济的"三驾马车".算法在促进经济社会发展的同时,也产生了算法黑箱、算法歧视、"大数据杀熟"、算法合谋垄断、算法奴役、"信息茧房"等新型社会问题.我国关于算法的立法规定散见于《电子商务法》《个人信息保护法》《反垄断法》等法律法规和部门规章之中,缺乏系统性和协调性,且对诸多问题未作规定.我国宜对算法产业促进与安全监管等进行专门立法,立法的主要内容包括总则、算法产业促进、算法安全监管、算法企业的权利与义务、算法救济、法律责任、附则等.本建议稿以算法安全与发展的平衡协调为基本理念,着眼于算法产业生态建设和系统治理,建立算法安全事前、事中、事后相结合的一体化监管体系,按照算法的应用场景、安全风险、社会影响等因素建立算法分类分级管理制度;确立监管沙盒、算法安全检测评估与认证、算法审计、算法公平性验证、算法追溯、算法终结、算法问责、算法救济、算法合规等安全保障机制,实时分析算法技术与应用的安全风险并动态采取预警和处置措施;针对算法黑箱等具体问题采用可信解释等解决方案,确保算法决策的科学性、合理性、可靠性;在规范算法研发与应用活动的同时引入产业政策扶持、容错免责及技术局限免责等鼓励创新的措施,推进算法在经济社会领域的深度融合与应用,以保护公民、法人和非法人组织的合法权益,维护国家安全和社会公共利益,探索并推动科学立法以维护算法公平与正义.
Law on the Algorithm Industry Promotion and Governance(Proposed Draft)
Algorithm is the core technology of artificial intelligence,and along with computing power and data,they are known as the"troika"of the intelligent era or digital economy.While algorithms promote economic and social develop-ment,they also generate new social problems such as algorithmic black boxes,algorithmic discrimination,big data-ena-bled price discrimination against existing customers,algorithmic collusion monopolies,algorithmic enslavement,and in-formation cocoons.The legislative provisions on algorithms in China are scattered in legislation such as the Electronic Commerce Law,Personal Information Protection Law,Anti-Monopoly Law,lacking systematicity and coordination and failing to provide for many issues.China should formulate specialized legislation on the promotion and security supervision of the algorithm industry,which mainly includes general provisions,algorithm industry promotion,algorithm security su-pervision,rights and obligations of algorithm enterprises,algorithm remedies,legal responsibilities,and supplementary provisions.This legislative proposal draft takes the balance and coordination between algorithm security and development as the basic concept.The draft focuses on the ecological construction and system governance of the algorithm industry,es-tablishes an integrated regulatory system that combines pre-supervision,in-process supervision,and post-supervision of algorithm security,and establishes a classification and hierarchical management system for algorithms based on factors such as application scenarios,security risks,and social impact of algorithms.It establishes a regulatory sandbox,algo-rithm security detection and evaluation and certification,algorithm audit algorithm fairness verification,algorithm tracea-bility,algorithm termination,algorithm accountability,algorithm relief,algorithm compliance and other security guarantee mechanisms,real-time analysis of security risks in algorithm technology and applications,and dynamic imple-mentation of warning and disposal measures.For specific problems such as algorithm black boxes,solutions such as trus-ted interpretation are adopted to ensure the scientific,reasonable,and reliable decision-making of algorithms.While reg-ulating algorithm research and application activities,the legislation should introduce innovative measures such as indus-trial policy support,fault tolerance exemption,and technical limitation exemption to promote the deep integration and ap-plication of algorithms in the economic and social fields,protect the legitimate rights and interests of citizens,legal per-sons,and unincorporated organizations,safeguard national security and social public interests,and explore the promotion of scientific legislation to maintain algorithm fairness and justice.

algorithm ethicsalgorithm black boxalgorithm discriminationalgorithm accountabilityalgorithmic relief

孙占利

展开 >

广东财经大学法学院,广东广州 510320

算法伦理 算法黑箱 算法歧视 算法问责 算法救济

国家社会科学基金

20BFX017

2024

重庆邮电大学学报(社会科学版)
重庆邮电大学

重庆邮电大学学报(社会科学版)

CHSSCD
影响因子:0.719
ISSN:1673-8268
年,卷(期):2024.36(1)