智能平權：通過數字開源的人工智能立法實現

智能平權是我國人工智能發展與治理的新要求。開源人工智能具有技術民主化、規模定律脫鉤性、監督開放化等宏觀價值，應以數字開源的人工智能立法實現智能平權，並設置相應的微觀法律保障措施以維護開源方的合法權益。在開源架構下，人工智能立法需要參考客觀語境與技術現狀，以數據分類分級保護機制預防開源人工智能潛在的數據安全風險，具體需對人工智能系統中的公共數據、商業數據以及個人數據提供不同保護模式，並根據數據內涵差異設置不同級別的保護強度。數字開源時代需要實現從部門性立法向領域性立法的轉型，以領域性立法框架實現智能平權的人工智能立法，即在橫向上優化自身與其他法規的銜接流程，在縱向上吸收人工智能專業知識以提升立法科學性。

Equitable Access to AI represents a new imperative for the development and governance of artificial intelligence in China. Open-source artificial intelligence offers macro-level benefits such as the democratization of technology, decoupling from scaling laws, and transparent oversight. Achieving Equitable Access to AI requires legislation aligned with digital open-source principles, along with corresponding micro-level legal safeguards to protect the legitimate rights and interests of open-source contributors. Within an open-source framework, AI legislation should reflect the actual technological context and conditions, incorporating data classification and tiered protection mechanisms to address potential data security risks. Different protection models should be established for public, commercial, and personal data in AI systems, with protection levels tailored to the nature of the data. The digital open-source era calls for a transition from sector-specific to domain-specific legislation. A domain-specific legislative framework for Equitable Access to AI should horizontally improve coordination with other laws and regulations, and vertically integrate specialized AI expertise to enhance the scientific rigor of legislation.