제 21 호 What Do You Think of “Open AI” Being Sued for Defamation and Copyright Infringement?
Kicker: OPINION
What Do You Think of “Open AI” Being Sued for Defamation and Copyright Infringement?
By Su-Young Kim, Reporter /kimsuyoung1342@gmail.com
Have you heard of ChatGPT? It is an interactive (Chat) artificial intelligence service based on Generative Pre-trained Transformer (GPT) developed by Open AI. These days, you can see many college students doing assignments using ChatGPT. Since ChatGPT was released, it has been attracting attention and has been used by many people. Recently, however, ChatGPT has been going through an unsavory affair, being embroiled in lawsuits such as defamation and copyright infringement. Let's find out what's going on and discuss what you think about issues like this.
According to Mark Walters, a gun rights activist in the U.S., ChatGPT described him as a “charge of embezzlement”. Walters heard from reporter Fred Riehl about ChatGPT's hallucinations. Liel asked ChatGPT to summarize Walters' nonprofit organization, the Second Amendment Foundation (SAF). However, ChatGPT accused Walters of embezzling SAF funds. It is a typical case of hallucination. Walters immediately filed a defamation lawsuit against Open AI. Open AI emphasized, "Chat GPT is not subject to defamation because it is not a publication." Open AI claimed, "Furthermore, its reputation was not flawed because no listeners actually believed in the content of malice or slander." For defamation requirements to be established in the United States, they must be false, must be announced to third parties without immunity, and the defendant's responsibility for intentional or negligence must be clear. In addition, the litigant must suffer substantial damage. The Georgia judge rejected the request to dismiss Open AI. Technica, a U.S. IT media outlet, said, "Open Ai’s claim does not appear to have been convincing to the judge," adding, "The judge ordered the trial to resume."
*Litigant: A person who requests the court to exercise jurisdiction in his or her name and the person subject thereto such as plaintiff and the accused
First, what is hallucination? It is a phenomenon in which false information is produced convincingly as if it were true. It is also the biggest drawback of generative AI. For this reason, companies hesitate to put generated AI into their systems or products. This is because if AI gives a wrong answer in the decision-making process within a company, it can have fatal consequences for management. Recently, AI hallucination solutions through search cognitive generation (RAG) have emerged. RAG is a technology that makes generative AI come up with a reasonable answer. Before AI answers, it searches for its own source or evidence. After that, the answer is presented based on the relevant evidence. Until now, the main methodology was the fine-tuning method, which retrains the model by putting corporate data into LLM.
In addition to Walter's defamation lawsuit, the U.S. daily New York Times filed a lawsuit on December 27, 2023, against open AI and Microsoft for copyright infringement. The New York Times filed a lawsuit in the Federal District Court in Manhattan, New York, arguing that "millions of articles published by the company were used to train chatbots as sources of information, and are now competing as reliable information providers." The New York Times negotiated with Microsoft and Open AI in April, but when it broke down, it filed a lawsuit. The lawsuit said it "must be held liable for billions of dollars in legal and actual damages" without exact financial demands.
Last year, the Korea Online Newspaper Association (KONA) said, "Artificial intelligence learning is a service that processes the original work at a qualitatively different level from the service stipulated in the existing news service terms and conditions," and "It is not only unfair to use news without consent based on the 'research purpose' clause of the old terms and conditions, but also a clear copyright infringement." Also, the Korea Association of Newspaper called on Big Tech to stop widespread copyright infringement, such as using news content without prior consent or source specification when developing generative artificial intelligence. It was a request for Big Tech to consult with copyrighted media companies in advance on the standards and methods of using news content for the development of generative artificial intelligence technology and to establish a reasonable compensation system. This demand started first in the foreign media. Digital Content Next (DCN), an organization affiliated with the media group, released a table to ensure intellectual property protection, transparency, accountability, fairness, and safety, saying, "Most of the actions of media companies using content for the purpose of learning or producing results go beyond the "fair use" allowed as an exception to copyright law violations." It is 7 Principles for Development and Governance of Generative AI. According to this principles, generative AI developers and distributors should respect copyright holders' rights to content, and media companies have the right to negotiate and fairly compensate for the use of intellectual property rights. Copyright law should stipulate that content creators are protected from unauthorized use of content, and generative AI systems should be transparently disclosed to media companies and users. Also, Generative AI system developers should be held accountable for the results, and there should be no risk that generative AI systems will cause or cause unfair market competition. Finally, generative AI systems should be safe and minimize the risk of privacy infringement.
What do you think? Do you think AI's hallucinations and copyright issues are serious? Or do you think it's too sensitive? How will you use Generative AI like ChatGPT in the future?
