OpenAI, the AI company led by Sam Altman, has agreed to allow access to its training data for the first time. This move is part of an ongoing legal battle with authors who claim their copyrighted works were used without permission to train the company’s AI models. On Tuesday, a group of authors suing OpenAI, including Sarah Silverman, Paul Tremblay, and Ta-Nehisi Coates, announced that they had reached an agreement on protocols for inspecting the company’s training data. The agreement allows OpenAI data to be inspected in copyright cases involving major authors. The authors aim to determine whether their works were used to train OpenAI’s technology without consent, a key issue in their copyright infringement lawsuit. The case could set significant precedents for the development of automated chatbots and other AI systems.
The authors initially filed a lawsuit against OpenAI, alleging that the company harvested vast quantities of books from the web to train its generative AI, ChatGPT. The lawsuit claims that OpenAI used these novels without providing credit or compensation, thus violating copyright laws. The authors argue that when ChatGPT generates answers or summaries of their works, it is producing infringing content.
Earlier this year, a court dismissed several claims against OpenAI, including accusations of unfair business practices. However, the authors’ direct copyright infringement claims have remained. U.S. District Judge Araceli Martínez-Olguín previously dismissed allegations of vicarious copyright infringement, negligence, and unjust enrichment, rejecting the idea that every output generated by ChatGPT could be considered a direct result of copyrighted materials.
The Data Inspection Agreement
OpenAI data is to be inspected in copyright cases and the inspection of will take place at the company’s San Francisco office. . Access will be granted in a highly controlled environment. The inspection will take place on a computer with no internet or network access, and all parties involved must sign non-disclosure agreements.
The use of electronic devices will be strictly limited. No cell phones, computers, or cameras will be allowed in the inspection room. OpenAI will allow note-taking on a computer provided in the room, and all notes will be transferred to another device under the supervision of OpenAI representatives at the end of each day. Copying of the training data itself into any notes is strictly prohibited.
OpenAI to Argue Fair Use
OpenAI has defended its practices by claiming that its training involves analyzing publicly available datasets, which may include copyrighted works. The company plans to argue that the use of these materials is protected under the legal doctrine of fair use. Fair use allows the use of copyrighted material for transformative purposes, such as in secondary works.
The authors, however, maintain that OpenAI trained its model using vast numbers of books obtained from shadow libraries. They allege that the AI was then able to generate summaries and detailed analyses of the themes in their novels, constituting copyright infringement.
Ongoing Legal Battles in AI
The OpenAI data, which is to be inspected in copyright cases may set a precedent for how AI companies handle copyrighted materials. This lawsuit is part of a wider series of legal challenges against AI companies, including Meta, over the use of copyrighted content to train AI models. Lawyers from the Joseph Saveri Law Firm are leading the legal action on behalf of the authors in both cases.
In a recent hearing, U.S. District Judge Vince Chhabria expressed concerns over the lawyers’ progress, pointing out that they had failed to conduct any depositions, which has raised doubts about the management of the case.
A request by the plaintiffs to take 35 depositions or 180 hours of testimony has been met with skepticism, as it was filed just 18 days before the scheduled close of fact discovery. Judge Thomas Hixson, who is overseeing the case, remarked that the timeline for conducting these depositions was “impossible” given the short timeframe remaining.
Also Read: Sam Altman’s Views on Psychedelic Drug That Made 8 CEOs Quit Revealed.
