Copyright: US Court Rejects Some of the Claims against OpenAI
A federal court in California recently dismissed some of the claims prominent figures in the entertainment industry have raised against OpenAI, the operator of the artificial intelligence platform ChatGPT. The plaintiffs lodged these claims as part of a copyright infringement lawsuit involving ChatGPT. This lawsuit is one of a series of legal proceedings brought against OpenAI, inter alia, by the Writers Guild of America and the New York Times, who alleged infringement of copyright in protected works.
Most of the rejected claims focus on the ChatGPT output’s alleged infringement of copyright in protected works. The court dismissed the allegations, and ruled that the claimants, including Sarah Silverman, could not demonstrate a substantial similarity between ChatGPT’s outputs and the protected works. Therefore, the court could not determine that infringement had occurred. The court moved on to reject the plaintiffs’ claims that if the judicial system did not hold OpenAI legally responsible, it would suppress artists and undermine the mechanism for encouraging creativity that is at the essence of copyright law. At the same time, the court rejected the claim that OpenAI unjustly enriched itself at the plaintiffs’ expense.
Not the Last Word
It is important to emphasize that the court’s ruling is far from the last word in the dispute. The Sword of Damocles still hangs over OpenAI’s head, and those of other artificial intelligence (AI) operators as well. For example, the plaintiffs’ allegation that the system has committed large-scale copyright infringements to train the models it uses still stands. The court has not yet decided on OpenAI’s main defense claim, which focuses on fair use. Additionally, the court has not yet addressed OpenAI’s claim that holding AI operators liable will hinder technological progress. The very fact that the court did not reject the infringement claim in its ruling may give some indication, at least, of its strength. Therefore, the legal threat to OpenAI is still very real.
Copyright and Agreements
The subjects of copyright and the protection thereof are core issues for companies that use AI technology. Therefore, it is important to make sure that the agreements in place between different players in the economy and AI technology providers include provisions that protect such players against legal claims in relation to copyright, whether with respect to the AI system training stage or with regard to its output.
We note that at the end of 2022, Israel’s Ministry of Justice published its position that using copyright-protected content to train artificial intelligence systems normally falls under lawful use arrangements, including fair use, and therefore does not constitute copyright infringement. It is important to clarify that Israeli courts have not yet examined this position and it is unclear if and how they will accept it. We expect interesting developments on this front as well.
***
Dr. Avishay Klein is the head of Barnea Jaffa Lande’s Privacy, Data Protection and Cyber Department.
Adv. Ran Karmi is an associate in the firm’s Antitrust and Competition Department.