OpenAI faces potential lawsuit from New York Times – Interesting Engineering

Credits: Jakub Porzycki/NurPhoto via Getty Images 
By subscribing, you agree to our and You may unsubscribe at any time.
The New York Times may sue OpenAI over its AI chatbot ChatGPT, which uses the newspaper’s stories to generate text, sources have told NPR. The paper is unhappy that OpenAI is not paying for the use of its content and is also worried that ChatGPT could reduce its online traffic by providing answers based on its reporting.
The sources, who spoke on condition of anonymity, said that the Times and OpenAI have been in talks for weeks to reach a licensing agreement, but the negotiations have turned sour and the paper is now exploring legal options.
A lawsuit from the Times would be a landmark case in the emerging field of generative AI, which creates new content from existing data. ChatGPT is one of the most advanced examples of this technology, as it can produce coherent and relevant text on various topics by using a huge dataset scraped from the internet.
However, this dataset includes millions of articles from the Times and other publishers, which were copied without permission. The Times claims that this violates its intellectual property rights and that ChatGPT is competing with its journalism by using its stories as a source of information.
The paper is especially concerned about the impact of ChatGPT on search engines, which are increasingly using generative AI tools to provide answers to users’ queries. Microsoft, which has invested heavily in OpenAI, is using ChatGPT to power its Bing search engine.
The Times fears that if users get answers from ChatGPT that are based on its reporting, they will have less incentive to visit its website and read its articles. This could hurt the paper’s revenue and reputation, said one of the sources.
The legal implications of generative AI are still unclear, as there is no precedent for such cases. However, experts say that the Times could have a strong case against OpenAI if it can prove that ChatGPT infringes on its copyrights.
According to US law, the court could order OpenAI to destroy its dataset and pay up to $150,000 for each infringement. This could be a huge blow for OpenAI, as it would have to rebuild its dataset from scratch using only authorized content.
“It’s a very serious issue for AI companies that use generative models,” said Daniel Gervais, an intellectual property expert at Vanderbilt University. “They need to be careful about what data they use and how they use it, or they could face legal consequences.”