Don't Believe The Tech Bro Hype: Copyright Owners Are The Real Winners In Kadrey v. Meta
The Judge Set A Very Big Copyright Trap For AI Developers Who Continue To Train On Unlicensed Content
Dear Readers, are you as exhausted as I am? This is my third “emergency” newsletter in three days. And this isn’t even my day job! I know it gets exhausting for you too — continuously bombarded with updates about literally everything. BUT — all of us reading this (my newsletter) do it because we need to know what’s going on in the world of AI as it relates to the world of media and entertainment. And what a week it’s been (and it’s only Wednesday as I write this).
It’s “Fair Use” Week!
First, we get the first “fair use” decision in the generative AI context — Bartz v. Anthropic (I wrote about it not once, but twice! Read this one). And now, this — Kadrey v. Meta (which makes it a “fair use” two-fer from the Northern District of California). Both cases involve individual book authors who sued Big Tech AI developers for, among other things, taking their individual books and adding them to their LLM training data sets — always without consent and compensation.
Look Behind The Kadrey Headline: This Time, Bury The Lead
Here’s what Silicon Valley headlines say about the new “fair use” decision in Kadrey v. Meta — AI developers won on the issue of AI training and fair use. Tech titans are likely high-fiving each other this morning, telling each other, “See, we can scrape copyrighted works with without consent and compensation — in other words, with impunity!”
But look beyond the hype and read Judge Vince Chhabria’s full legal opinion, young Skywalker (like I did). Only then will you see the light and become the true Jedi master who knows WTF he/she is talking about (because the Tech Bros who claim victory have no idea about what just hit them).
The Judge’s Decision Is An Extremely Narrow & “Painful” One
‘Tis true that Judge Chabbria handed Meta a "fair use" victory on the issue of AI training on unlicensed copyrighted works (books). But the Judge bemoans — over and over again — that his hands were tied under the specific facts and legal arguments presented to him. He takes great pains to emphasize that his decision is an extremely narrow one, applying only to the 13 individual authors/plaintiffs in his court. These are his exact words: “And, as should now be clear, this ruling does not stand for the proposition that Meta’s use of copyrighted materials to train its language models is lawful. It only stands for the proposition that these plaintiffs made the wrong arguments and failed to develop a record in support of the right one.”
Read those words again. They are truly remarkable. I’ll translate his “legalese” for you. “Plaintiffs, you f’d up with your lawyers!” Yes, Judge Chhabria went there — essentially blaming the author/plaintiffs' lawyers for their narrow loss (not the merits of the case).
Judge Chhabria Just Laid Out The Roadmap
So if the plaintiffs’ lawyers made the wrong arguments, what were the right ones? “Market Substitution” and “Market Dilution” — core concepts of Factor 4 of the relevant 4-part “fair use” test — are what the Judge had ordered. But, according to Chhabria, plantiffs’ lawyers presented no evidence on that dispositive issue. If they had, the Judge very clearly states that he would have ruled the other way.
Here’s how the Judge himself words the “winning argument” that plaintiffs should have made:
“The plaintiff might argue that, even if the model can’t regurgitate their own works or generate substantially similar ones, it can generate works that are similar enough (in subject matter or genre) that they will compete with the originals and thereby indirectly substitute for them.”
The Judge then goes on to say this:
“Harm from this form of competition is the harm of market dilution. Or as one commentator describes it, the harm of ‘indirect’ substitution, rather than ‘direct’ substitution.”
So here’s the reality of what this ground-breaking legal decision and opinion represents. It’s a Manifesto of sorts, an easy-to-follow roadmap, that establishes — for the first time (in a way Bartz v. Anthropic only wishes it would have) — clear rules of the game that describe when AI developers must pay copyright owners for using their creative works as food for their AI training. And the bottom line answer is this — “MOST OF THE TIME!”
The Judge’s Conclusion? AI Developers Must Pay Most of the Time!
As I waded page-by-page through Judge Chhabria’s thoughtful legal opinion and analysis — an opinion that is worth reading in full because it tackles the “meta” issues (so to speak) of AI and the creative process (including artist incentives) — my smile grew ever wider. I found myself smiling in solidarity with my Creative brothers and sisters who have been steamrolled by Big Tech Silicon Valley players over and over again through the years, all in the predictable and now-lazy refrain of “progress.”
But “progress” doesn’t mean “pay less.” Remember, Big Tech reportedly is investing $1 Trillion into AI. So the money most certainly is there to recognize the value of the Creative Community’s contributions to AI’s LLM core value proposition.
Judge Chhabria makes this point time and time again, downright chiding AI developers for using scare tactics as a defense to the simple act of engaging in straightforward business licensing transactions. These are just some of the Judge’s mocking words:
“Meta seems to imply that such a ruling [i.e., rejecting fair use'] would stop the development of LLMs and other generative AI technologies in its tracks. This is nonsense.” He then writes this: “So where copying for LLM training isn’t fair use, LLM developers (including Meta) won’t need to stop using copyrighted works to train their models. They will need only to pay rightsholders for licenses for that training.”
Amen Brother Chhabria! Preach!
So, Tech’s knee-jerk “win” is objectively nothing but a mega long-term loss for the AI developer community’s “fair use” agenda. Yes, Judge Chhabria — like Judge William Alsup the day before — is just one federal district court judge amongst many who will tackle this AI training “fair use” issue. But early decisions in novel cases carry out-sized long-term precedential impact. And if I were a betting man, Judge Chhabria’s more deeply reasoned and thoughtful decision will carry more weight than his judicial brother Judge Alsup’s.
Judge Chhabria Emphasizes That AI Copying Is Not The Same Thing As Human “Copying”
Speaking of Judge Chhabria, here in Kadrey v. Meta, in a remarkable paragraph of his written opinion, he flatly rejects the reasoning of his colleague Judge Alsup who decided his “fair use” issue only one day earlier in the same federal district court. Chhabria cites Alsup’s own written opinion and says this about it:
“Judge Alsup focused heavily on the transformative nature of generative AI while brushing aside concerns about the harm it can inflict on the market for the works it gets trained on. Such harm would be no different, he reasoned, than the harm caused by using the works for ‘training schoolchildren to write well,’ which could ‘result in an explosion of competing works.’ According to Judge Alsup, this ‘is not the kind of competitive or creative displacement that concerns the Copyright Act.’ But when it comes to market effects, using books to teach children to write is not remotely like using books to create a product that a single individual could employ to generate countless competing works with a miniscule fraction of the time and creativity it would otherwise take. This inapt analogy is not a basis for blowing off the most important factor in the fair use analysis.”
Wow! Judge Chhabria, how do you really feel?
But of course he’s right. AI copying is not the same as human “copying” for creative inspiration. I’ve been saying the same thing forever, including in my post yesterday about Bartz v. Anthropic in which I lay out why Judge Alsup’s reasoning is flat out wrong.
Expect The AI Licensing Market To Heat Up: AI Developers, Judge Chhabria Puts You On Notice
So, expect the AI content licensing market to heat up. Significantly. The federal district courts of California’s Northern District — in back-to-back days — just gave AI developers twin “head fake” hollow victories, only to snatch them away with searing long-term defeats.
ATTENTION ALL AI DEVELOPERS! You are now officially on notice that those of you who continue to “take” without consent and compensation (you know, “theft,” in the words of Disney v. Midjourney) in order to create AI tools intended to substitute and dilute markets, will be forced to pay, one way or another. One such threatened market, Judge Chhabria noted, is news-gathering. Do I hear The New York Times v. OpenAI, anyone? It is 100% clear to me that Judge Chhabria would reject OpenAI’s “fair use” defense and award significant and potentially massive damages to The Times if that case were in his court.
Here’s my bottom line advise, dear Tech community (of which I’ve been a part for decades now). Choose to cooperate, rather than litigate. It will be far less expensive — and far more productive — for you.
[What do you think? Send me your feedback to peter@creativemedia.biz.]
About Peter & His Firm Creative Media
Peter is a leading AI, media and entertainment expert who is a Harvard Law educated IP lawyer and business entrepreneur who has negotiated $4+ billion in deals over the course of his career. His firm Creative Media represents media companies and rights-holders in AI content licensing (both for training and for RAG-fueled outputs/displays), with deep relationships and market insights and intelligence second to none. Peter and his firm also advise on AI strategies and overall execution. He is well-known by the key players inside AI tech and can reach key decision-makers and influencers in record time to execute. Not just talk.
More generally, Peter and Creative Media specialize in breakthrough business development and M&A and cost-effective legal services in the worlds of media, entertainment, AI and tech.
Reach out to Peter at peter@creativemedia.biz to explore working with us.
Send your feedback to me and my newsletter via peter@creativemedia.biz.
Thanks for unpacking this. Solid, nuanced “look under the hood” of this decision.