It has been reported that the Latest York Times (NYT), which OpenAI sued for copyright infringement, charged an enormous fee to make use of ChatGPT to search out evidence to prove the damage. Because of this, NYT needed to incur additional costs for model inspection in addition to litigation costs.
In a copyright infringement lawsuit filed by NYT on the fifteenth (local time), Arstechnica claimed that OpenAI is attempting to make a profit by charging retail prices as ChatGPT usage fees.
Based on this, following the court’s motion, the NYT was in a position to inspect the Open AI model and find parts that copied the authors’ articles.
Nevertheless, during this process, Open AI provided support for investigation cooperation.
This included that the NYT could hire an authority to examine ChatGPT data in a secure location designated by OpenAI in an environment without access to the Web or other networks.
Moreover, the variety of queries that experts can check on ChatGPT through the API is restricted to credits price $15,000 (roughly 20 million won) based on retail price. If this limit is exceeded, the NYT requires users to pay half the retail price for the service. The logic is that since NYT’s activities use the traffic vital for Open AI’s normal business activities, it must pay corresponding costs.
Nevertheless, the NYT claimed that OpenAI was attempting to hide the infringement by citing excessive and unspecified costs. As well as, it was revealed that with the intention to secure the vital evidence, credit price about $800,000 (roughly KRW 1.1 billion) was needed, which is greater than 50 times the quantity presented by OpenAI. As well as, in a document submitted to the court, it was criticized that “Open AI hid the associated fee of operating actual traffic and only presented the quantity charged as a service to general firms.”
In response to this, OpenAI explained that setting an initial limit was a vital measure to stop NYT’s unreasonable demands. Open AI responded by saying, “NYT is attempting to run an arbitrary, unfounded and unnecessary search on the Open AI model, and we demand that Open AI bear all costs.”
Specifically, it was revealed that the knowledge presented to the NYT this time was just like the knowledge provided in other lawsuits against the writers.
Meanwhile, experts imagine that this controversy could have a big impact on future AI copyright infringement lawsuits.
If the court determines that OpenAI can charge retail prices for model testing, plaintiffs who cannot afford to rent AI experts or conduct model testing at business prices will likely have difficulty filing lawsuits.
NYT and OpenAI declined to comment on the lawsuit.
Reporter Park Chan cpark@aitimes.com
