OpenAI relies on publicly available data and human users’ feedback for reinforcement learning to generate human-like interactions with its diverse ChatGPT users, aiding them in streamlining their workflow and providing them with personalized assistance. While OpenAI claims that it uses such data in a “manner protected by fair use and related principles,” Raw Story Media Inc. and AlterNet Media Inc. brought forth a lawsuit in February claiming otherwise. More specifically, the Plaintiffs claim that OpenAI is wrongfully using thousands of articles without permission to train its AI and generate copyrighted responses for its users. The Plaintiffs ultimately assert that OpenAI’s removal of the articles’ copyright management information (CMI) is illegal. Despite the novel and pertinent issue raised, Raw Story and AlterNet Media’s suit was dismissed for lacking sufficient standing.
The Digital Millennium Copyright Act and Standing
Congress passed the Digital Millennium Copyright Act (DMCA) in 1998 to address existing gaps between copyright and the Internet. Three key provisions of the DMCA include: “(1) establishing protections for online service providers in certain situations if their users engage in copyright infringement, which allows copyright owners to inform online service providers about infringing material so it can be taken down; (2) encouraging copyright owners to give greater access to their works in digital formats by providing them with legal protections against unauthorized access to their works (for example, hacking passwords or circumventing encryption); and (3) making it unlawful to provide false CMI (for example, names of authors and copyright owners, titles of works) or to remove or alter that type of information in certain circumstances.”
In their complaint, Raw Story Media and AlterNet Media claim that OpenAI violated the third provision of the DMCA by removing CMI—authorship and title—from articles used to train ChatGPT. The Plaintiffs claim that OpenAI had the choice to train ChatGPT “using works of journalism with copyright management information protected by the DMCA intact” but opted to strip it away. In doing so, the Plaintiffs claim that OpenAI trained ChatGPT “not to acknowledge or respect copyright, not to notify ChatGPT users when the responses they received were protected by journalists’ copyrights, and not to provide attribution when using the works of human journalists.”
OpenAI responded to the complaint, arguing that the Plaintiffs lack sufficient standing to bring forth the lawsuit. Standing concerns a litigant’s right to have a court rule upon the merits of a particular claim for which they seek judicial relief and is rooted in Article III of the Constitution. To establish standing, the Supreme Court holds that a party must allege a genuine stake in the outcome of a case because they have personally suffered: (1) a concrete and particularized injury; (2) that is traceable to the allegedly unlawful actions of the opposing party; and (3) that is redressable by a favorable judicial decision.
Judge Colleen McMahon wrote in a decision issued in the United States District Court for the Southern District of New York. In her decision, Judge McMahon ruled in favor of OpenAI, stating that “removing CMI (author names) to train generative AI tools, absent evidence of dissemination of those works, does not qualify as an adverse effect needed to establish standing.” The decision denied the Plaintiffs’ request for injunctive relief. The Plaintiffs alleged there was a substantial risk that ChatGPT could reproduce the articles verbatim without credit and claimed injunctive relief to compel OpenAI to remove its content from its training data. Finding this argument speculative, Judge McMahon held that the Plaintiffs did not plausibly allege that this “substantial risk” was “certainly impending.”
Judge McMahon further denied the Plaintiffs’ motion for leave to replead without prejudice and granted OpenAI’s motion to dismiss. Though this dismissal was without prejudice, granting Plaintiffs a narrow window to amend their complaint, Judge McMahon casts doubt on the possibility of a favorable ruling for the Plaintiffs: “I am skeptical about Plaintiffs’ ability to allege a cognizable injury.”
Through this decision, the United States District Court for the Southern District of New York sent a message to current and future adversaries of OpenAI. The court demands unshakable evidence of concrete and particularized injuries, not mere allegations.
Business Implications of Immunizing OpenAI
This suit is among many claims brought forth against OpenAI and other tech companies. Many other writers, media outlets, creators, and artists have been similarly aggrieved from AI compromising the protection of their creative output. Despite Judge McMahon’s dismissal, some remain optimistic about such future cases, believing that further evidence of infringement claims can overcome the current shortcomings of DMCA claims. While the ruling restricted the use of DMCA CMI removal as a method for regulating AI training, some believe that alternative methods demonstrating actual harm and real-world impacts can address the rising copyright issue. The Intercept, a media group pursuing similar legal action against OpenAI regarding DMCA claims, is among the optimistic, seeking to limit OpenAI. Having reportedly gathered 600 pages of exhibits exemplifying OpenAI’s capacity to produce recreations of other’s work, Intercept’s lawsuit has the potential to overcome shortcomings present in Raw Story’s case. To be ruled on later this month, Intercept’s lawsuit demonstrates at least one case demonstrating ChatGPT’s identical replication of its articles. However, critics believe that DMCA will not permit such claims.
As of now, OpenAI has not commented on its plans to change its model training techniques, maintaining that its practices align with existing legal frameworks and copyright laws. In the absence of change, a different outcome, or alternative avenues for restricting AI, Raw Story’s dismissal and the theory of “no standing” can significantly reduce creators’ ability to protect their work in model training. It can additionally limit cases for copyright infringement. As Professor James Grimmelmann warns, “This theory of no standing is actually a potential earthquake far beyond AI. It has the potential to significantly restrict the kinds of IP cases that federal courts can hear.” Given the plethora of lawsuits against OpenAI and other similar tech companies, time and future court rulings will shed more light on this concern and clear the direction of future regulation.
*The views expressed in this article do not represent the views of Santa Clara University.
Comments