Apple has an AI plan that ignores marketing noise

Billions of dollars are being invested in the generative AI gamble, and companies have made sure the numbers are big enough to be featured in media publications around the world. In addition to monitoring AI developments, newsrooms are a big part of the AI ​​industry.

The technology companies developing these AI models, such as OpenAI, Google, and even Apple, have an insatiable appetite for contracts that give them exclusive access to the copyrighted information in their language models. We are tying up with the big guys in a hurry.

ChatGPT maker OpenAI recently made headlines for partnering with Axel Springer, the parent company of Politico and Business Insider. The Information also reports that OpenAI is offering between $1 million and $5 million a year in licensing fees for copyrighted news articles to train AI models. This is one of the first signs of how much AI companies are willing to pay for licensed content.

Meanwhile, in the Cupertino orchard, Apple is also shaking trees for media collaborators, dangling at least $50 million over multiple years for the privilege of accessing data. The list appears to be a portrait of media royalty, with Condé Nast, publisher of Vogue and The New Yorker, receiving a call from Cupertino. NBC News, owner of People, The Daily Beast, and Better Homes and Gardens, is another news company approached by IAC.

Apple’s No Noise AI

Apple’s negotiations with news publishers may seem like an early attempt to catch up in the field of generative AI, which allows machines to work and speak like humans. But some may disagree, as the iPhone maker isn’t relying on AI as a marketing claim. The trillion-dollar company has been quietly developing AI capabilities that haven’t received as much media attention as other companies.

Plus, secrecy is like a religion at Apple. It’s deeply embedded in the company’s DNA and culture. The company is notorious for keeping quiet about new product announcements and releases.

For example, in October 2023, Apple and Columbia University jointly released an open source multimodal language model called Ferret. Initially, the paper received little attention, but the buzz increased after a few weeks due to community interest in the possibility of his local LLM powering small devices.

The Cupertino-based company also introduced two new research papers describing new techniques for 3D avatars and efficient language model inference. This could enable more immersive visual experiences and enable complex AI systems to run on consumer devices such as iPhones and iPads.

First, the researchers introduced HUGS (Human Gaussian Splats), a tool that converts videos shot with a single camera into 3D avatars. In the second paper, we discovered how to use large language models such as GPT-4 on regular devices without using too much memory.

Existing functionality

The unexpected news of Apple’s move into open source and local ML development comes at a time when companies are focusing on hardware capabilities. The designer of GPT, OpenAI, is likely working on a smartphone designed by his famous iPhone designer Jony Ive.

Siri, Apple’s virtual assistant, has been largely stagnant since its release, but given its existing capabilities, it’s expected to be reimagined next with generative AI. Apple’s CEO Tim Cook said Apple’s AI efforts are “in progress,” but otherwise declined to elaborate publicly.

While Apple is busy courting media companies, it’s being secretive about how it plans to bring generative AI to the news industry. Apple has an army of news-loving people glued to its devices. Some news executives are hopeful that Apple’s move could amount to a substantive partnership. Several insiders also expressed optimism about the deal’s long-term potential.