比較法視閾下我國生成式人工智能立法模式研究

Improve the Legislative Approach to Generative Artificial Intelligence

生成式人工智能立法供給及其水平事關國家科技安全、資源配置效能及數字經濟發展質量，完備的法律體系能夠為生成式人工智能技術發展提供有力支撐和制度保障。良好的立法可以避免國家科技安全風險或企業知識產權涉訴風險，推動我國人工智能技術跨越發展、人工智能產業優化升級，加快形成新質生產力。全球規範生成式人工智能技術發展的立法主要有“全面監管式”“激勵創新式”和“新發展主義式”三種模式。我國現階段可採激勵創新為主、全面監管為輔的策略，在實踐中豐富中國特色新發展主義的意蘊內涵，平衡科技安全、技術創新與產業發展的關係，發揮我國人工智能產品或服務的市場應用潛力。建議制定專門《生成式人工智能法》，堅持以人為本、新發展主義、可信任問責及綠色原則四大基本原則，重點調整利用生成式人工智能創作的內容權屬；結合應用場景對風險分級分類，劃分生成式人工智能服務者前端注意義務及後端備案審查義務，配套完備的侵權責任認定標準，不斷提升我國人工智能領域的立法水平和國際規則制定的話語權，從而加速實現生成式人工智能技術全面趕超。

The legislation and level of generative artificial intelligence supply are crucial to national technological security, resource allocation efficiency, and the quality of digital economic development. A comprehensive legal system can provide strong support and institutional guarantees for the development of generative artificial intelligence technology. It can avoid national technological security risks or corporate intellectual property litigation risks, promote the leapfrog development of artificial intelligence technology and the optimization and upgrading of the artificial intelligence industry, and accelerate the formation of new quality productive forces. The legislation for the development of global standard-generating artificial intelligence technology mainly has three models: ''comprehensive supervision'', ''incentive innovation'' and ''new developmentalism''. At this stage, China can adopt the strategy of incentivization innovation and comprehensive supervision, enrich the connotation of new developmentalism with Chinese characteristics in practice, balance the relationship between scientific and technological security, technological innovation and industrial development, and give play to the market application potential of China's artificial intelligence products or services. It is suggested to formulate a special Generative Artificial Intelligence Law, adhere to the four basic principles of people-oriented, new developmentalism, trustworthiness and accountability and green principle, and focus on adjusting the content ownership created by using generative artificial intelligence; Combined with application scenarios, the risk classification is divided into front-end care obligations and back-end filing review obligations of generative artificial intelligence service providers, complete supporting standards for the identification of tort liability, and constantly improve the legislative level of China's artificial intelligence field and the discourse power of international rule-making, so as to accelerate the comprehensive catch-up of generative artificial intelligence technology. Give play to the institutional influence of AI governance with Chinese characteristics.