The New York Times Co.’s claiming suit, asserting that the AI chatbot, ChatGPT, has been reproducing almost verbatim text from its published articles, has caused repercussions throughout the AI industry. As per the details of the complaint, this dispute is pivotal to OpenAI Inc. and its partners, who are on the verge of the most notable copyright battle they’ve faced so far.
The 70-pages long complaint, filed in a Manhattan federal court, states that OpenAI Inc., in collaboration with Microsoft Corp., scrapped the Times’s articles without a proper license. The preliminary text and URL of the original article are utilized by the chatbot to output text that mirrors the published work. If the Times manages to win this lawsuit, in addition to damages amounting to billions of dollars, OpenAI might be also ordered to eliminate any models and training sets incorporating the Times’ articles.
The laudable Kristelia García, a copyright law professor at Georgetown University, explains the Times’s view of ChatGPT drawing exact facts from their articles or mimicking the writing style. A wide variety of lawyers specialized in copyright and technology issues agree on the perilous legal position OpenAI has taken.
The legal battle is not only confined to copyright issues but also to the Times’s recent attempts to strike a licensing deal with OpenAI. The recent arrangement between OpenAI the German publishing giant Axel Springer SE, owner of Politico and Business Insider, is a testament to the ramifications of this case.
OpenAI’s defenses may also be undermined by the licensing market for news content. This lawsuit, brought by a standalone company, brings into focus OpenAI’s argument that the copying of published works falls under the “fair use” doctrine.
One particularly illustrative exhibit enclosed with the complaint provides comparisons, clearly demonstrating the almost identical output of OpenAI’s most advanced model, ChatGPT-4, derived from Times articles. This evidence furnishes a compelling argument showing the chatbot’s capability of memorizing the training data.
The final verdict will depend hugely on the fair use doctrine under copyright law. If OpenAI is found guilty of market harm due to copying Times’ articles and making them available for free, it could lead to significant implications on future copyright cases involving AI models.
Although OpenAI has claimed that it is committed to respecting the rights of content creators, this ongoing legal battle raises some serious questions about the appropriate use of AI in re-creating copyrighted work. The final outcome of this lawsuit will undeniably shape future conduct and precautions taken by AI companies with regard to intellectual property rights.
Learn more about this case here.