Apple Wants to Partner With Major Publishers to Train AI

[MacRumors]

Apple has approached several major publishers to establish deals that would allow the Cupertino company to train generative artificial intelligence systems on news content, reports The New York Times.

Apple is aiming for multiyear deals and has approached Condé Nast, NBC News, and IAC. Condé Nast publications include Vogue, Wired, Vanity Fair, Ars Technica, Glamour, The New Yorker, GQ, and more, while IAC owns publications like People, The Spruce, Serious Eats, Martha Stewart Living, Real Simple, Entertainment Weekly, and Better Homes & Gardens.

Proposed deals have been worth at least $50 million, and would allow Apple to license archives of news articles. According to The New York Times, some of the publishers were "lukewarm" on Apple's offer. Apple's terms are said to be "too expansive," and Apple has been vague about how it will apply generative AI to news.

Other publishers were "optimistic" about a potential partnership, and were pleased that Apple asked to use their content rather than just training generative models on published news without permission as other AI companies have done.

Multiple rumors have suggested that Apple is working overtime to catch up to its rivals on generative AI offerings, with Apple testing an "AppleGPT" chatbot internally and planning for new AI features in iOS 18.

Microsoft, Google, and Meta have all incorporated generative AI into their products over the course of the last year, which means Apple is lagging behind when it comes to AI technology. ChatGPT, the most popular chatbot from OpenAI, was trained on a huge amount of data that includes books, articles, and web pages.

In addition to the copyright issues that come with the broad scraping of internet content, ChatGPT has sometimes been criticized for the accuracy of the information that it sometimes surfaces. By training an AI model on a more tailored set of information, Apple could end up with a more reliable product. Apple is also said to be planning to incorporate generative AI features across its app offerings, so a model that has been fed content from news sources could perhaps be added into Apple News.

The New York Times says that Apple executives have been "debating" how to get the data needed for generative AI products. Apple has not wanted to source information from the internet because of its focus on privacy, so deals with news publishers provide an alternative.

Article Link: Apple Wants to Partner With Major Publishers to Train AI

 

[kildraik]

Good grief. Talk about choosing completely unbiased and factually reliable publication groups to train data on.

 

[Flowstates]

This is deep entertainment (species of publications), the need for un-biasedness is lessened by essence and the required subjectivity to inherent to this kind of information "News" it is not.

 

[Hogswarts]

Even if unsuccessful, such talks may serve to slow down the competition.

 

[CarAnalogy]

Oh no don’t do that. Siri is dumb and pop culture obsessed enough as it is.

 

[canadianreader]

I'm not sure why wouldn't they train their bots on 2 or more sides of the same story. I still think open-source chatbots are more balanced than agenda driven big corporations chatbots.

 

[now i see it]

End users (us) couldn’t give a rat’s ass about ai.
It’s all techno-nerd hype.

 

[wigby]

Good grief. Talk about choosing completely unbiased and factually reliable publication groups to train data on.

I don't think any of these publications listed are trying to be unbiased. They're just glorified opinion pieces for readers of a certain political persuasion and of course, Apple already has relationships with many of them via their News app. But that begs the question...Who is unbiased and factual in the publishing space?

 

[Arislan]

Good grief. Talk about choosing completely unbiased and factually reliable publication groups to train data on.

Maybe they could train a second AI off of Truth Social. See how that works out. LOL

 

[kildraik]

Maybe they could train a second AI off of Truth Social. See how that works out. LOL

That would be just as bad, or worse, much Like CNN/Warner Media Group would be, and is luckily not on this list.

But that begs the question...Who is unbiased and factual in the publishing space?

Large firms? Likely none. Independent agencies? Sure, but good luck letting Big Tech scrape their data willingly.

 

[Baritone_Guy]

Good grief. Talk about choosing completely unbiased and factually reliable publication groups to train data on.

It’s only a start. I am sure there will be others added with less factual and unbiased sources in the future.

 

[iBluetooth]

Good grief. Talk about choosing completely unbiased and factually reliable publication groups to train data on.

Is it any worse than using Reddit or X? This is though the first time that I hear about someone not just scraping the data, without asking to pay.

 

[Baritone_Guy]

End users (us) couldn’t give a rat’s ass about ai.
It’s all techno-nerd hype.

Ignore at your own peril.

 

[arc of the universe]

this is excellent news.

generative language models are like children.
children, like leaves, dont fall far from their parental tree trunk.

its clear that the results of the dozen or so of each of the emerging LLM giants relies on the data bases they have used to shape the output.

diverse cultures (read: foreign cultures) need to be included in the original data bases. because, what can (and probably will) emerge, are culturally specific results (LLM shaped by largely usamerican input will bend towards a usamerican bias; LLM that used western european data bases will yield a western european bias).

generative AI LLM represent a once in a 1000 years opportunity to help the world become closer together instead of building walls between people.

great literature and thorough journalism can come from many many countries. usa, eqypt, russia, china, japan, greece, italy, south africa. everywhere really.

people use a combination of deductive reasoning and inferential inductive reasoning; objective experience and learned personal biases. all of us on this planet are living results of these biases, received from our parents and from our cultures.

i hope apple partners with many many non-usa publishers as well.
as John lennon told us, Imagine..

 

[IllegitimateValor]

AI TERFBot trained on NYTimes bigotry and antique racism incoming…

 

[DocMultimedia]

I'm sure Apple will ask MR for permission to use these forums to train their AI. What could possibly be better than the clarity around here. Lol

 

[WarmWinterHat]

I'm sure Apple will ask MR for permission to use these forums to train their AI. What could possibly be better than the clarity around here. Lol

There's a 50-50 chance that the AI would end up believing it's a God, or it would commit suicide.

 

[5232152]

Good grief. Talk about choosing completely unbiased and factually reliable publication groups to train data on.

What do you want? Trumps biography as the source of truth?

 

[Plutonius]

Oh no don’t do that. Siri is dumb and pop culture obsessed enough as it is.

I can see Apple replacing Siri with a true AI but it may not work as well as they hope .

 

[sjsharksfan12]

Maybe these companies can train AI not to kill us all in a nuclear attack in 30 or so years (see Terminator 2).

 

[kildraik]

Is it any worse than using Reddit or X? This is though the first time that I hear about someone not just scraping the data, without asking to pay.

Perhaps, perhaps not. Output would ultimately rely on the guardrails in place. But the partnerships they are seeking are a good indicator of information and it’s biases they wish to provide the end user.

 

[bluecoast]

Great that they are doing this and wanting to pay for the material.

I’m sure that they’ve approached way more publishers than the report states too.

 

[Student of Life]

Who is unbiased and factual in the publishing space?

Humans by nature are biased. I guess the better question is who is less biased and again that depends on the individual. To some Fox News is the most accurate news organization, to others it’s pure trash. To some CNN is more accurate and honest, while to others it’s the spawn of satan. It’s hard to answer.

 

[Surf Monkey]

I’d hope that these media companies, which will absolutely sell their IP for scraping sooner or later, flip the script on Apple and demand a high monthly subscription fee for it.

 

[GrantNeufeld]

It’s encouraging to me that Apple is taking a “get permission” and compensate the copyright holders approach, rather than the “we’ll just vacuum up everything, without consent, and without attribution or compensation” method used for the big Large Language Models (like GPT) that we have seen so far.

And let’s not assume that they would in any way limit their sources to the few examples listed in this article. As a global company, I would expect they would probably be drawing on resources from around the world.