Apple to Spend $50 Million to Train Its AI from News Sources

While rumors of an “AppleGPT” tool have been swirling for months, the iPhone maker has generally been pretty cagey about what direction it plans to go with generative AI.

Apple CEO Tim Cook has made no secret of Apple’s AI ambitions; however, when pressed, Apple’s chief executive repeatedly points to its products and features, emphasizing that AI and machine learning are “fundamental technologies” that are at the heart of everything from Personal Voice and Live Voicemail to Crash Detection.

Nevertheless, the hints are there that Apple is at least considering something in the conversational AI space that could be similar to ChatGPT. We all know that Siri has a lot of room for improvement, and Apple is reportedly spending billions of dollars on AI servers and machine-learning-focused research and development. Sources also suggest the iPhone 16 will get a better mic for Siri’s new AI powers that are rumored to be coming next year in iOS 18.

Now, it seems Apple is ramping up its efforts to train its generative AI system. A new report in The New York Times reveals that it’s begun negotiating with “major news and publishing organizations” to get access to their vaults of material to feed its new AI bot.

The Times heard from “four people familiar with the discussions” who said that Apple is trying to line up multi-year licensing deals with several high-profile publishers that could reach upwards of $50 million to open up their collective archives.

The technology giant has floated multiyear deals worth at least $50 million to license the archives of news articles, said the people with knowledge of talks, who spoke on the condition of anonymity to discuss sensitive negotiations. The news organizations contacted by Apple include Condé Nast, publisher of Vogue and The New Yorker; NBC News; and IAC, which owns People, The Daily Beast and Better Homes and Gardens.

The New York Times

Not unlike human intelligence, generative AI systems need to be “taught” to recognize patterns that can build the neural networks that allow them to do their thing.

Other companies like Google have been able to leverage the large volume of data that it collects from its search engine and other products and customers, but Apple doesn’t have those kinds of resources and has strong privacy policies that tie its own hands when it comes to the customer data that it does have stored in iCloud.

In other words, Apple would rather pay $50 million (which, let’s face it, is couch change to Apple) to feed its AI from external sources than try to siphon data from its customers. In fact, collecting customer data is so anathema to Apple that it won’t even collect analytics from social media sites like Twitter/X because some of its customers might post on the site.

Of course, legitimate news publications also tend to provide more stable and reliable material to work with since they’re meticulously curated, edited, and accurately sourced. There’s no word on the type of material Apple is looking for, but based on the list of publications, it would presumably include everything from text material to photos.

Since negotiations are at an early stage, it’s also hard to say whether Apple will have much success. The Times reports that some publishers are “lukewarm” to the idea after being burnt by the likes of Facebook and Meta. There are also concerns that Apple is looking for overly “expansive” terms that include passing the liability for Apple’s use of the content back onto the publishers and being unclear on how generative AI would be used in potentially competing apps and services such as Apple News+.

However, some executives remain optimistic and are reportedly impressed with Apple’s approach of seeking permission first rather than forgiveness later — something that many other AI companies have been accused of, only seeking licensing deals after they’d already snarfed up a massive amount of content from news organizations.

[The information provided in this article has NOT been confirmed by Apple and may be speculation. Provided details may not be factual. Take all rumors, tech or otherwise, with a grain of salt.]