OpenAI and Apple are hurrying to reach agreements with the press to gain access to copyrighted information for AI models.
Billions of dollars have been thrown at the generative AI wager, and the businesses made sure the amounts were large enough to make it into global media journals. Apart from keeping an eye on AI advances, newsrooms have become an increasingly important aspect of the AI sector.
With an insatiable thirst, tech giants like OpenAI, Google, and even Apple building these AI models are scrambling to make deals with newspaper moguls to exclusively access protected information for their language models.
OpenAI, the maker of ChatGPT, has recently made headlines for partnering with Axel Springer, the parent company of Politico and Business Insider.
According to The Information, OpenAI pays between $1 million and $5 million annually to license copyrighted news stories to train its AI models. This is one of the first indications of how much money AI firms are willing to spend on licensed content.
Meanwhile, Apple is shaking the trees in Cupertino in search of media collaborators, offering at least $50 million over a multiyear period for the privilege of accessing data. Condé Nast, owner of Vogue and The New Yorker, is among the names on the list, which reads like a who’s who of media aristocracy. NBC News has been approached, as has IAC, the owner of People, The Daily Beast, and Better Homes & Gardens.
No Noise AI from Apple
The talks between Apple and news organizations may appear to be an early attempt to catch up with generative AI, which allows machines to work and communicate like people. However, one can assert that Apple is not using AI as a marketing ploy. The trillion-dollar corporation has been quietly developing its AI capabilities, which have not received as much public attention as the others.
Furthermore, confidentiality is almost a religion at Apple. It’s profoundly ingrained in the company’s DNA and culture. It has a reputation for keeping new product announcements and launches under wraps.
For example, in October 2023, Apple and Columbia University collaborated to release Ferret, an open-source multimodal language model. The paper garnered little attention at first, but curiosity about the possibility of local LLMs to power small devices surged a few weeks later.
The Cupertino-based corporation also released two new research papers containing new methodologies for 3-D avatars and fast language model inference, which might enable more immersive visual experiences and allow complicated AI systems to direct on customer devices such as iPads and iPhones.
First, the researchers presented HUGS (Human Gaussian Splats), a technique for converting single-camera films into 3D avatars. They discovered a way to employ huge language models, such as GPT-4, on normal devices without requiring too much RAM in the second study.
Capabilities Already in Place
Apple’s surprising entry into open source and local ML research comes at a time when corporations are focusing on their hardware capabilities. OpenAI, the GPT engineer, is most likely working on a smartphone created by renowned iPhone designer Jony Ive.
Given its current capabilities, Apple’s virtual assistant, Siri, which has remained relatively stagnant since its release, can be expected to be rebuilt using generative AI in the near future. Tim Cook, Apple’s CEO, has stated that the company’s AI work is “ongoing,” but has not provided any additional details publicly.
While Apple is constantly courting media outlets, they are keeping its plans for generative AI in the news sector under wraps. Apple has a sizable news-loving audience hooked to their smartphones. Some news executives are optimistic that Apple’s move will result in a real-world relationship. A couple of insiders are also optimistic about the deal’s long-term prospects.
