- Apple is offering major news organizations and publishers at least $50 million to get access to their content for training the generative AI model it’s developing
- Sources say Apple is brokering a deal with Conde Nast, publishers of Vogue and the New Yorker; IAC, publishers of People and the Daily Beast; and NBC News for the purpose
- Apple is currently testing an in-house ChatGPT-style AI chatbot dubbed “Apple GPT” that will be announced for customers later next year
The New York Times reported on Friday that Apple has opened negotiations with several major publishers and news organizations, seeking permission to use their content in training the generative AI model it’s developing.
Apple has reportedly floated the idea of multi-year deals worth at least $50 million to license the archives of news articles, cited NYT sources familiar with the matter.
Apple Seeks Archived Content From Major Publishers to Train its ChatGPT-Style AI
The new organizations that the iPhone and Mac maker got in touch with include Conde Nast – publishers of Vogue and the New Yorker, NBC News, and IAC – which owns People, the Daily Beast, and Better Homes and Gardens.
While the prospective deal is enticing, some publishers are concerned about the repercussions of letting Apple use their content throughout the years. They believe that a broad licensing deal for their archive could lead to legal issues along the way. They are also worried about the potential competition that may arise from Apple’s initiative.
However, Apple has been building goodwill in that regard by asking the publishers permission to use their content and showing a willingness to compensate for their efforts. According to NYT, company executives have been in discussion for years about how to gather data for developing its generative AI system.
The move will also be a departure from Apple’s usual strategy of collecting privacy-centric data. The tech giant has been reluctant to use information sourced from the internet due to its commitment to protecting users’ privacy.
Publishers Concerned About Apple’s Move Following Copyright Lawsuits Filed Against AI Companies
While Apple has been rather quiet in the AI space, firms like OpenAI, Google, Microsoft, and Adobe have developed generative AI tools for their various products and services. Furthermore, these firms have also been accused by creators of infringing copyrights and stealing their content to use it to train their respective models.
Microsoft-backed AI startup OpenAI is currently fighting several lawsuits filed by novelists, including George R.R. Martin and John Grisham. A group of non-fiction authors who also sued Microsoft and OpenAI said both companies have built a business “valued into the tens of billions of dollars by taking the combined works of humanity without permission”.
Meanwhile, the only area Apple has revealed using a transformer model is in the overhauled autocorrect feature in iOS 17 for the iPhone.
Apple Reportedly Working on “Apple GPT”
Bloomberg’s Mark Gruman reported that Apple has been internally testing a generative AI chatbot called “Apple GPT” that uses the company’s “Ajax” framework, which is designed to act as a base for its various machine learning projects.
Apple has already deployed AI-related advancements to its search, Siri, and maps based on Ajax. The framework is now being used to create large language models and serve as the foundation for a ChatGPT-style tool used internally.
Gruman also noted that the system still requires special approval for access to contents and Apple has asserted that any output from the chatbot can’t be used to develop features bound for its customers.
Apple employees are reportedly using the model to assist with product prototyping. Additionally, the system also summarizes texts and answers questions based on data it has been trained with.