Apple reportedly wants to develop its own Large Language Model (LLM) and needs training content for this. The reports the New York Times. To ensure the quality is right and copyrights are protected, Apple wants to negotiate multi-year contracts with publishers totaling at least $50 million to access their news archives. The goal would be to use large amounts of real news content to train and improve Apple’s AI models.
Publishers Apple has contacted include household names such as Condé Nast, NBC News and IAC/InterActiveCorp. However, it is claimed that some publishers have so far remained skeptical of the proposal.
If the licensing deals are successful, Apple could gain access to the archives of popular publications such as Vogue, The New Yorker, The Daily Beast and People.
Compared to competitors like Google and Microsoft, Apple has so far held back when it comes to AI. However, licensing of publisher content could indicate increased investment in this area.
Apple’s AI ambitions
Apple is rumored to have an internal AI chatbot called Apple GPT, which is similar to ChatGPT. The company also introduced AI to improve autocorrect in iOS 17. Further, more advanced AI functions could follow.
Licensing news archives could allow Apple to train generative AI models on real data. This could lead to smarter and more natural AI conversations and address a problem that other LLMs certainly have: the source of the training content is often the result of crawlers, which, like search engines, read content from websites without having a license for reuse . Meanwhile, if Apple can guarantee that its LLM contains only copyright-free content, that could be a big plus.
Apple is following the example of other AI players. OpenAI has already signed agreements with the Associated Press and Axel Springer to use their news content for AI training purposes.