From The New Publishing Standard:
Apple has been slow to jump on the AI bandwagon in any meaningful way, which has given rivals a head-start, but also let them make the mistakes and the public-relations fails so Apple can jump in now looking like the good guy.
“Apple has reached out to news entities like Condé Nast, the publisher of Vogue and the New Yorker and more.”
. . . .
For western businesses, Christmas is a good time to bury controversial news, so no surprise Apple waited until the publishing industry was in “Holidays” limbo before making this announcement.
Per Business Today, Apple has put $50 million on the table to entice news publishers to give permission to use their content to train Apple’s own gen-AI system.
Apple has been slow to jump on the AI bandwagon in any meaningful way, which has given rivals a head-start, but also let them make the mistakes and the public-relations fails so Apple can jump in now looking like the good guy.
If Apple can be the first do to a major deal with content providers it will certainly take the moral high ground, but beyond that there’s little in the reports emerging so far to suggest Apple has anything original to offer.
Link to the rest at The New Publishing Standard
PG wondered when he read the announcement how much of the $50 million will be going to the authors who wrote the news items.
While he has never seen a contract between a newspaper and one of its writers, he would bet quite a bit of money that the news writers didn’t sign any sort of contract that allowed the publisher to utilize their work to train an AI system.
If any visitors to TPV have a copy of an employment agreement between a newspaper and one of its writers, PG would be interested in reviewing it. Click the Contact PG button to initiate a discussion.
Speaking of publishing wanting free money:
https://www.windowscentral.com/microsoft/new-york-times-sues-microsoft-and-openai
At least they openly admit it is about money and not principle.
“The New York Times has announced it is suing OpenAI and Microsoft over AI use of copyrighted work. Breaking the cardinal rule of a news outlet of never allowing yourself to become the story, the New York Times has decided to stand up to Microsoft and OpenAI in the hopes of helping their bottom line.
In the lawsuit filed by the New York Times, the plaintiff discusses how the use of New York Times copyrighted material by Microsoft and OpenAI is monetarily impacting the New York Times. While I personally agree that the issue at hand here needs to be addressed, the fact that I can’t actually read the New York Times coverage on this lawsuit on their own website because I would have to subscribe to read it is likely a larger reason for their declining revenue. ”
More at the source, who have their own heartburn at Chatbot answering in bruef questions covered by entire articles of theirs, despite links to those same articles.
In a comment to the source, it is pointed out that the loss of revenue they bemoan might have “just a bit” to do with the paper’s editorial bias and questionable decisions on which viewpoints are allowable in opinion pieces (the Hamas-appointed mayor of Gaza) and which are firing offenses (allowing a piece by a sitting US Senator.)
Popcorn time!
A reminder, Microsoft has extended indemnification coverage to users of its AI tools:
https://news.bloomberglaw.com/artificial-intelligence/how-risky-is-microsofts-pledge-to-defend-ai-copyright-lawsuits
“Microsoft built features into the models designed to reduce purely duplicative outputs that could raise copyright concerns—for example, natural language filters that “try to limit the number of verbatim outputs that we can include in any particular response,” the Microsoft lawyer said.
Microsoft’s indemnification commitment “puts their money where their algorithmic development is,” said Chris Callison-Burch, an associate professor of computer and information science at the University of Pennsylvania’s School of Engineering and Applied Science who studies AI. It’s a good move for PR and reassurance, he added, which “also aligns with their technical goals of trying to mitigate this unintentional replication of copyrighted data.”
Microsoft has been ready for the court fight for months.
“If Apple can be the first do to a major deal with content providers it will certainly take the moral high ground, but beyond that there’s little in the reports emerging so far to suggest Apple has anything original to offer.”
Considering how fast the generative software explosion is moving, if Apple is starting now, they will be behind the curve indefinitely because not only is the software evolving in “internet time” it is simultaneously diversifying and spawning dozens if not hundreds of specific vertical applications via small language models.
Apple’s MO of waiting for somebody else to pioneer a market and then march in with their own version and hundreds of millions in marketing won’t work with “AI” because “AI” isn’t a singular market or product but a software development approach more akin to object-oriented programing or, more recently, sandboxing. As such people won’t be buying AI per se but rather APPs that incorpoate generative functions internally, typically as upgrades to established products. Especially in vertical markets and enterprise software, two categories where Apple has near zero presence. So they’ll not only be showing up late to the party, they’ll be showing up with zero installed base.
Best bet is the “winners” in AI software will be the existing category leaders: Adobe, Oracle, Google, Microsoft, Facebook, Autocad, Nuance, etc. There will be some jockeying for position but it is doubtful there will be much market share shifting. To paraphrase the old BASF ads: “AI won’t make new software but it will make old software better.”
Where Apple can be a player is in consumer hardware, along with Google, Qualcom, AMD, NVIDIA, and Intel.
But again, they’re late there.
And people aren’t going to switch from Android or Windows to Apple over “AI”. No consumer hardware NPU looks to be significantly better than any other anytime soon.
So, realistically, it makes no difference how Apple trains whatever model they cook up.
Oh, and those $50M?
Pocket change to anybody outside publishing.
And a waste once the court verdicts roll out.
Here’s a good look at Amazon’s AWS AI server hardware:
https://m.youtube.com/watch?v=njyZR0Ceke0&pp=ygUad3NqIHRlY2ggbmV3cyBhbWF6b24gY2hpcGQ%3D
There is mention of everybody that matters in AI servers.
Notice who isn’t anywhere to be found.
Apple talking about generative AI is akin to GM talking about getting into business jets. Theoretically possible (Honda did it) but of zero importance until they actually do it and *succeed* at it. Anything but that is just puffery. And publishing thinking that if Apple throws them a bone they can leverage the real players into giving them free money. Yeah, that’s going to happen.
(Remember when the BPHs were going to leverage their copyrights to get Hollywood to guve them billions? Plenty of talk, no real change.)
Corporate publishing is a stagnant business and pining for free money from other businesses is just wishful thinking by their cheerleaders.
I can’t speak to “newspapers,” which tend to be — what’s the word? oh, yeah, rapacious — regarding the rights they claim to “need” ever since Tasini and Muchnick. (And if you don’t agree to those terms, you get blacklisted.)
But according to Condé Nast boilerplate from {date withheld to protect the guilty, but at a time that derivative uses like training of large-language-model-based systmes was foreseeable}, it’s not even close… for works not explicitly commissioned as works made for hire. But that’s only as to freelancers; employees are, in technical terms, screwed because they’re employees. Past experience with other major periodicals houses indicates that the publishers have even less claim on the right to create derivative works from freelancers’ contributions.