According to media reports, the U.S. District Court for the Southern District of New York recently dismissed a copyright lawsuit filed by news outlets Alternet and Raw Story against OpenAI. However, this victory may be temporary. The court's ruling did not address the core issue at the heart of the most controversial debates in the AI field: whether the use of copyrighted content to train AI models requires authorization.
The two media outlets filed the lawsuit in February of this year, accusing OpenAI of removing copyright management information (CMI) during the training data process, including authors' names, usage terms, and titles of works. The lawsuit sought a minimum of $2,500 in damages for each infringement and requested the court to prohibit OpenAI from continuing to use their copyrighted works.
The main reason the court dismissed the lawsuit was that the plaintiffs failed to prove specific damages suffered due to the removal of copyright information. OpenAI argued in its defense that the plaintiffs could neither prove that ChatGPT was trained on their works nor provide evidence of specific losses. The judge agreed with this view and noted that given the scale of the database, the likelihood of ChatGPT outputting the plaintiffs' articles was small.
You Yunting, a senior partner at Shanghai Daban Law Firm, stated that proof has always been the key challenge in AI copyright disputes. Due to the "black box" nature of large models, it is difficult to prove whether specific works were used for training. Under the current legal framework, there is a lack of systems to assist the disadvantaged party in obtaining evidence.
Currently, OpenAI faces at least six related lawsuits, including those from The New York Times, Daily News, and a collective lawsuit from a group of writers. All these cases revolve around a core issue: whether AI companies need authorization to use copyrighted content for training models.
It is worth noting that countries have different attitudes towards this issue. Japan has classified the use of copyrighted works for AI training as "fair use," while courts in China and the United States have not yet provided a clear answer. Professor Yao Zhiwei from the School of Law at Guangdong University of Finance and Economics pointed out that the theory of fair use lacks legislative basis in China, and its judicial recognition is highly uncertain.
Although this lawsuit was dismissed, the judge indicated in the judgment that the plaintiffs could refile the lawsuit regarding the issue of OpenAI using their works for AI training without paying for them. The attorneys representing Raw Story and AlterNet have already stated that they will amend the complaint to continue defending their rights.