The U.S. District Court for the Northern District of California heard a case concerning whether copyrighted content can be used in AI model training.Make a major rulingThe court ruled for the first time that such use falls within the scope of "fair use," providing a legal defense for AI companies like Anthropic. This ruling not only attracted attention from the industry but also caused great disappointment among creators who have long worried about the unauthorized use of their works by AI.
This case was filed in 2024 by several writers such as Andrea Bartz, Charles Graeber and Kirk Wallace Johnson, accusing the artificial intelligence company Anthropic of unauthorized crawling and using the content of their published books to train large language models (LLMs), constituting copyright infringement.
However, William Alsup, the judge who presided over the case, pointed out after the trial that although the content used by Anthropic may have come from unauthorized sources, the purpose and nature of its use for training artificial intelligence models complies with the principle of fair use, and such use is legally transformative and therefore does not constitute direct infringement.
The ruling stated that even if Anthropic initially used so-called "pirated e-books" and subsequently purchased them, while not completely excluding its liability, the statutory damages amount could be adjusted accordingly. In other words, while the court acknowledged the legitimacy of the use of copyrighted content in the development of artificial intelligence, it also did not completely abandon the possibility of pursuing legal action for illegally obtained content.
This ruling marks the first time that the U.S. court system has made a substantive ruling on the source of generative AI training data. It may also provide a certain degree of reference for other AI technology developers, especially in the development stage where generative AI models are highly dependent on large-scale data sets for training.
It is worth noting that although the court did not completely deny the rights of creators, the overall trend is relatively favorable to the artificial intelligence industry.
This result is expected to trigger more discussions on the ethics of creation and technology. As the application scope of generative artificial intelligence continues to expand, everything from text to images and music can be quickly generated through models, and the commercial challenges and risks to the value of creators' works will become increasingly severe.
Similar lawsuits are likely to continue to increase in the future, and are expected to further push legislatures to establish clearer regulations regarding the sources of artificial intelligence data, authorization mechanisms, and compensation systems.
