OpenAI, the organization behind ChatGPT and its large language models (LLMs) GPT-3.5 and GPT-4, is facing two copyright lawsuits filed by U.S. authors and comedian and actor Sarah Silverman. The plaintiffs claim that OpenAI used their copyrighted materials in the training data for its AI models. In response, OpenAI has filed motions to dismiss five out of the six counts in the lawsuits. The company defends the transformative nature of its LLM technology and argues for a balance between copyright protection and technological advancement.
OpenAI emphasizes the value and potential of AI, specifically ChatGPT, in enhancing productivity, aiding in coding, and simplifying daily tasks. The company sees ChatGPT’s impact as a revolutionary advancement, comparing it to the invention of the printing press. OpenAI believes that copyright has limitations and does not extend to facts and ideas. Processing copyrighted works and producing summaries or derivative works should not infringe copyright, according to the company.
OpenAI bases its defense on the fundamental facts of LLM technology, stating that it is a neural network trained on extensive text data to effectively comprehend human language. Users can input text prompts and receive generated content. OpenAI claims that its products, like ChatGPT, merge LLMs with parameters that ensure accuracy, relevance, safety, and utility of the generated outputs. The company argues that fair use exceptions in copyright law can accommodate transformative innovations like LLMs. It cites court cases where fair use doctrines protected the innovative use of copyrighted materials.
OpenAI challenges the claims of direct copyright infringement made by the plaintiffs. The company argues that fair use should not be addressed in a motion to dismiss since it is an affirmative defense that the defendant must prove. OpenAI calls for the dismissal of secondary claims, such as vicarious copyright infringement, violations of the Digital Millennium Copyright Act (DMCA), violations of California’s Unfair Competition Law (UCL), negligence, and unjust enrichment. The company presents evidence refuting the plaintiffs’ theories and contends that LLMs do not produce derivative works. It also challenges the claims about copyright management information being removed.
OpenAI finds deficiencies in the negligence and unjust enrichment claims, arguing that intentional acts negate any grounds for negligence. The company states that it did not hold on to profits or benefits from the infringed material. Additionally, OpenAI argues that both negligence and unjust enrichment claims are preempted by federal copyright law. It believes the state law claims should be dismissed based on legal reasoning.
While the immediate effects of the motion to dismiss may be limited, the outcome of the lawsuits could have significant implications for AI use cases and copyright law. The determination of whether training large language models on copyrighted works infringes copyright will shape the future of technological progress. OpenAI’s defense sheds light on the intersection of copyright law and AI technology, highlighting the transformative nature of AI and the need for a balanced approach.
OpenAI’s motions to dismiss the copyright lawsuits demonstrate the company’s commitment to defending its LLM technology and advancing AI. The organization argues for a balanced approach to copyright protection and technological innovation. As the lawsuits unfold, the legal conflict between OpenAI and the plaintiffs will likely shape the future of copyright law and determine the extent to which AI models can be trained on copyrighted materials. The outcome of these cases has the potential to redefine the boundaries of copyright law and pave the way for further advancements in AI technology.
Leave a Reply