OpenAI is racing against time to negotiate authorization with publishers and spare no effort to pave the way for GPT Store.

Zhitong App has learned that OpenAI, the developer of OpenAI, is accelerating negotiations with dozens of publishers to reach agreements to license their exclusive content or articles for OpenAI's data training. This is a broader licensing effort than previously known, as the AI startup is seeking more content to train its artificial intelligence models and is making every effort to embrace the "App Store moment" - the launch of the OpenAI Store, a custom application store based on its text generation AI model (such as OpenAI-4), which will be launched next week. "We are in negotiations and discussions with many publishers. They are very active and very positive, and progress is going well in my opinion," said Tom Rubin, OpenAI's Head of Intellectual Property and Content, in a media interview. "You have seen these announced deals, and there will be more deals in the future." A source familiar with the matter previously told the media that OpenAI recently signed a multi-year licensing agreement worth tens of millions of dollars with Politico's parent company, Axel Springer SE. In July 2023, OpenAI announced a significant agreement with The Associated Press, but did not disclose the specific amount of the agreement. These deals are crucial to OpenAI's future as the company balances the need for updates and accurate data to build its models, as well as increasingly stringent guidelines for data sources. However, last week, The New York Times, one of the media outlets that had negotiated with OpenAI, sued OpenAI and Microsoft, a major shareholder of OpenAI, for unauthorized use of the magazine's articles. This lawsuit poses significant challenges to OpenAI's multiple businesses. If The New York Times wins the lawsuit, OpenAI may not only owe billions of dollars, but may also be forced to destroy all of its training data, including data related to The New York Times' content, which is an expensive and complex task. However, more directly, this lawsuit complicates OpenAI's efforts to collaborate with the media industry. But Rubin said, "The current situation is very different from what publishers faced in the past with search engines and social media. Here, the content is used to train models, not to replicate content, let alone replace original content." However, The New York Times disagrees with OpenAI's position, believing that OpenAI has completely plagiarized its journalists' works without payment. In the lawsuit, the publisher presented examples of OpenAI copying entire passages of text from The New York Times almost word for word - some journalists even pointed out that in some cases, it specifically prompted OpenAI to replicate The New York Times' content. The publisher believes that this proves that OpenAI has illegally used The New York Times' data. The New York Times stated in a statement: "If Microsoft and OpenAI want to use our achievements for commercial purposes, the law requires them to obtain our permission first." "However, they did not do so." OpenAI's "App Store Moment" is coming: OpenAI App Store will be launched next week. If OpenAI fails to reach cooperation agreements with The New York Times and more news publishers, it will not only face huge compensation, but also face significant obstacles to its multiple businesses due to significant content gaps in data training, especially the upcoming "OpenAI Store". The so-called OpenAI Store allows users to participate in the development of custom applications similar to OpenAI and share and promote them. When others use the products, they can make profits, which is equivalent to an "App Store" based on OpenAI's large-scale model technology. According to media reports, OpenAI plans to launch the OpenAI Store, a custom OpenAI application store based on its text generation AI model (such as OpenAI-4), at some point in the next week. The OpenAI Store can be seen as an "APP Store" that combines OpenAI's large-scale model. The OpenAI Store aims to allow all users to share and sell OpenAI applications, such as chatbots based on OpenAI's large-scale model, customized for different purposes. Developing OpenAI applications in the OpenAI Store does not require programming experience, so ordinary participants can develop applications like professional developers. Developers only need to input the desired functions provided by the OpenAI application in simple language, and OpenAI Builder, OpenAI's application building tool, will try to create an AI-driven chatbot to perform these functions. During the process of building OpenAI applications, developers only need to submit simple dialogue instructions, additional knowledge data, and then choose whether they need multimodal functions such as web search, data analysis, and image generation, to achieve rapid development of specific fields such as law, medicine, and architecture, similar to OpenAI applications. For example, by completing some simple pre-operations, an OpenAI application can be trained on a recipe collection, so that it can answer questions about specific recipe ingredients, giving birth to an OpenAI application focused on the recipe field. Alternatively, OpenAI can absorb proprietary code libraries from specific companies so that more professional developers can review their application styles or generate code based on best practices. OpenAI states that developers building OpenAI applications must review the company's updated usage policies and OpenAI brand guidelines to ensure that their OpenAI applications comply with the requirements before they are eligible to be publicly launched in the OpenAI Store. They also need to verify their users' personal information and ensure that their OpenAI releases are "public" in nature.