OpenAI has announced a new partnership with Condé Nast. The deal will see the AI startup integrate content from the publisher’s leading brands, including GQ, Vogue and The New Yorker, into ChatGPT, its new search engine prototype SearchGPT, and similar products. Announced last month, SearchGPT combines the AI startup’s conversational models with web-sourced information – much of which is being directly piped into the model through partnerships with major publishers.
OpenAI’s chief operating officer Brad Lightcap hailed the firm’s deal with Condé Nast as the latest example of how SearchGPT and its other products will be augmented by ground truth daily supplied by trusted media brands. “We’re committed to working with Condé Nast and other news publishers to ensure that as AI plays a larger role in news discovery and delivery, it maintains accuracy, integrity, and respect for quality reporting,” said Lightcap.
OpenAI on campaign to win over major publishers
Its new partnership with OpenAI will see Condé Nast join several other publishers, including Associated Press, Axel Springer, the Financial Times, The Atlantic, Le Monde, and Vox Media in feeding article data directly into the startup’s AI products and services. “This partnership recognises that the exceptional content produced by Condé Nast and our many titles cannot be replaced,” said Condé Nast’s chief executive, Roger Lynch. It was, he added, “a step toward making sure our technology-enabled future is one that is created responsibly.”
Not every established media brand has greeted the appearance of its content on OpenAI’s products. In December 2023, the New York Times sued the startup for copyright infringement, claiming that OpenAI used its articles to train ChatGPT without its consent. Three months later the publication was joined by The Intercept, AlterNet and Raw Story, which all claimed that ChatGPT produced “verbatim or nearly verbatim copyright-protected works of journalism without providing author, title, copyright or terms of use information contained in those works.”
Neither is OpenAI the only household name in AI feeling the heat from copyright lawsuits. This week, Anthropic was hit by a class action lawsuit by authors and journalists Charles Graeber, Andrea Bartz and Kirk Wallace Johnson. In a complaint filed in a California court, all three claim that Anthropic used their works to train its flagship model, Claude, without their consent. “It is no exaggeration to say that Anthropic’s model seeks to profit from strip-mining the human expression and ingenuity behind each one of those works,” they said.
Fine-tuning GPT-4o
In a separate announcement, OpenAI has launched fine-tuning capabilities for its advanced large language model (LLM) GPT-4o. This new feature will allow developers to tailor GPT-4o using custom datasets, enhancing performance and reducing costs for specific use cases. According to OpenAI, fine-tuning offers the ability to adjust the model’s response structure and tone or to adhere to complex domain-specific instructions.
Developers are expected to achieve significant results with as few as a few dozen examples in their training datasets. This capability is expected to substantially improve model performance across diverse fields, from coding to creative writing.
Read more: Microsoft abandons observer seat on OpenAI board
