The U.S. government has submitted a statement to the court regarding the copyright case in which The New York Times sued OpenAI, publicly supporting this ChatGPT developer for using copyrighted materials to train large models without authorization. This move is not a judgment, but it demonstrates the clear stance of the Trump administration on copyright disputes related to AI training.
Submit 20 pages of opinion letters
According to TechCrunch, this 20-page court document states that the United States has a strong interest in continuing to develop a "powerful and competitive" artificial intelligence industry and hopes to maintain a dominant position in global AI practices. The document also cites an executive order signed by Trump last year, emphasizing that the U.S. needs to maintain its global leadership in the field of AI.
The focus of the case is whether AI Company can use books, news reports, and other copyrighted materials for model training without permission. Multiple publishing institutions, including The New York Times, have consistently argued that such practices constitute copyright infringement.
The controversy focuses on fair use.
At the heart of this legal debate is whether the relevant training activities fall under the category of "fair use" as defined by U.S. copyright law. Within this framework, courts typically assess whether the use in question possesses sufficient "transformative character," meaning whether the original content is being utilized for a different purpose rather than merely replicating its value.
The U.S. government stated in its opinion letter that restricting the development of large models due to a misunderstanding of the principle of fair use could hinder creativity and scientific research progress, as well as weaken the economic vitality of the United States.
In previous cases, most of them were favorable to the AI company.
Currently, regarding the multiple cases of copyright infringement related to AI training, the overall trend is relatively favorable for AI company. Reports mention that last year, judge William Alsup ordered Anthropic to pay $1.5 billion in copyright settlements to a group of writers, but the focus of the penalty was not on model training itself.
The issues identified by the court at that time were mainly that Anthropic used an illegal "shadow library" to obtain the books required for training, rather than simply using these works for training by the model itself. Alsup also compared large model training to human reading of works followed by creation in the judgment, suggesting that the models were not intended to directly copy or replace the original works.
May have an impact on the outcome of the case
TechCrunch pointed out that this opinion letter itself does not constitute a judicial ruling, as the case is still being heard by the United States District Court for the Southern District of New York, and the government department that submitted the opinion does not make judgments directly. However, the public intervention of the federal government at this time may still have an impact on the discussion of the case and its subsequent proceedings.
This also means that the debate in the United States regarding the legality of training data surrounding AI is escalating from a copyright dispute between companies and publishing institutions to an issue at the level of industrial competition and national policy.
