中圖分類號(hào)：D923.8;TP18文獻(xiàn)標(biāo)識(shí)碼：ADOI:10.19358/j.issn.2097-1788.2025.01.017引用格式：劉婷. 人工智能數(shù)據(jù)訓(xùn)練作品使用的困境破解——以不同法律主體地位為語(yǔ)境［J］.網(wǎng)絡(luò)安全與數(shù)據(jù)治理，2025，44（1）：104-109.

Cracking the difficulties in the use of artificial intelligence works——In the context of different legal subject statuses

Abstract： How to regulate the unauthorized use of works by artificial intelligence for training is an urgent issue that needs to be addressed in current copyright law. Given that different theories of legal subject status can lead to different interest distribution objects and affect the adjustment logic of rules, different interest relationships should be discussed in specific contexts.. If the "subject theory" is adopted, the relationship between infringement in the learning stage and infringement in the output stage should be clarified, and the standard of creative use and non creative use should be adopted to determine the outer edge of the reasonable use system. If the "tool theory" is adopted, artificial intelligence should bear heavier copyright obligations and stricter fair use access standards while referring to the application of fair use. With the help of collective management organization system, simplifying the licensing process and reducing licensing costs can help artificial intelligence achieve access standards.

Key words : artificial intelligence; legal subject status; use of works; reasonable use