AI訓練資料庫使用版權作品衍生多起訴訟
OpenAI公司開發的人工智慧聊天機器人程式ChatGPT， Meta Platforms公司開發的人工智慧語言模型LLaMA，以及Stability AI公司，皆使用大量的文學作品來訓練公司開發的人工智慧軟體。
普立茲獎小說獎得主 Michael Chabon、東尼獎最佳話劇獎得主David Henry Hwang、作家 Matthew Klam、作家 Rachel Louise Snyder 和作家 Ayelet Waldman的文學作品是「高品質長篇寫作的最佳範例」，而OpenAI公司、Meta公司以及Stability AI公司未經過這些著作權人的同意，就複製並使用他們的文學作品來訓練人工智慧軟體，這些作者針對這些科技公司侵害著作權的行為提起訴訟。
科技公司則認為複製並使用這些文學作品來訓練公司開發的人工智慧軟體，屬於著作權法的合理使用，並未侵害版權人對作品的版權。
依美國著作權法第107條的規定，要判斷科技公司未經版權人同意就使用版權作品來訓練人工智慧軟體是否為合理使用，必須考慮「使用的目的和性質」、「有版權作品的性質」、「同整個有版權作品相比所使用的內容和數量」、「這種使用對有版權作品的潛在市場或價值所產生的影響」此四項因素，為綜合判斷。
參考：
Multiple Lawsuits Initiated Over the Use of Copyrighted Works
in AI Training Datasets
The artificial intelligence chatbot program ChatGPT, as well as the large language models LLaMA and StableChat, developed respectively by OpenAI, Meta Platforms, and Stability AI, are all trained using a large number of literary works.
In particular, the literary creations of authors Michael Chabon, winner of a Pulitzer Prize for fiction, and David Henry Hwang, recipient of a Tony Award for best play, alongside with the works of authors Matthew Klam, Rachel Louise Snyder, and Ayelet Waldman, regarded as “prime examples of high-quality long-form writing,” have been copied and used to train artificial intelligence programs by OpenAI, Meta, and Stability AI, without the consent of their rightful copyright owners. In response, the authors filed copyright infringement lawsuits against the implicated technology companies.
The defendants contend that replicating and utilizing these literary works to train their artificial intelligence programs can be regarded as a “fair use” under copyright law, and argue that such actions do not infringe on the copyrights held by the plaintiffs in these works.
Pursuant to Article 107 of the U.S. Copyright Law, in order to determine whether the technology companies' unauthorized use of copyrighted works for training artificial intelligence programs qualifies as a “fair use,” four factors must be considered comprehensively, namely “the purpose and character of the use,” “the nature of the copyrighted work,” “the amount and substantiality of the portion used in relation to the copyrighted work as a whole,” and “the effect of the use upon the potential market for or value of the copyrighted work.”
Reference Materials:
- Reuters. (September 11, 2023). More Writers Sue OpenAI for Copyright Infringement Over AI Training. Available at: More Writers Sue OpenAI for Copyright Infringement Over AI Training (usnews.com) (Accessed: November 9, 2023);
- Reuters. (September 12, 2023). Pulitzer Winner Chabon, Other Authors Sue Meta Over AI Program. Available at: Pulitzer Winner Chabon, Other Authors Sue Meta Over AI Program (usnews.com) (Accessed: November 9, 2023);
- Copyright Law, 17 U.S.C. § 107 (1976). Available at: 17 U.S. Code § 107 - Limitations on exclusive rights: Fair use | U.S. Code | US Law | LII / Legal Information Institute (cornell.edu) (Accessed: November 9, 2023).
Statement: This article is an original work of authorship, and all referenced documents, data, photos, and materials have been listed in the "Reference Materials" section located at the bottom of the article.
by 智財研究部