Newsom's Hollywood/Silicon Valley GenAI Balancing Act (Someone's Gotta Do It, Since The Feds Won't)
Cali's 17 New GenAI Laws May Become the Nation's De Facto Standards
Time for your weekly brAIn dump! First, it’s the week’s “big stor-AI-s” (news you can use). Next, it’s the “showc-AI-se” — your GenAI “video of the week” (this one comes from award-winning filmmaker Paul Trillo) and “podcast of the week” (mine, created and voiced by Google’s remarkable new NotebookLM). Then, this week’s “mAIn event” (my feature story about Newsom’s past two frenetic — and arguably schizophrenic — weeks of new game-changing GenAI legislation). Finally, the “AI Litigation Tracker” (updates on the key AI infringement cases), now brought to you by the law firm McKool Smith (just named the nation’s “Plaintiff IP Firm of the Year” — with more “Top 100 Verdicts” than any other firm). McKool is on board to make the “AI Litigation Tracker” more useful than ever before.
There’s a lot in this week’s newsletter. So LET’S GO!!!!
I. The Big Stor-AI-s …
(1) OpenAI Swims in New C-AI-sh! (& MSFT Is Still Its Swim Coach)
OpenAI, the leading generative AI company on the planet, just announced an eye-popping $6.6 BILLION of new financing (at an even more eye-exploding valuation of $157 billion!) in its quest to be the first to attain artificial general intelligence (AGI). Microsoft continues to plow new money into the company — $750 million, on top of its already $13 billion — and Nvidia joined this round (even as CEO Sam Altman’s leadership team continues its mass exodus). Read more about it all here. Microsoft previously invested at a “mere” $29 billion valuation. I predicted then, in this newsletter, that $29 billion would be seen as a bargain in retrospect. I’ll take the “W” on that one.
(2) Meta Makes Mega Generative AI Video Moves
Mark Zuckerberg sent many Hollywood players reeling last week when Meta announced its OpenAI Sora “killer,” Movie Gen. Movie Gen isn’t available yet, but Meta is naturally saying it’s the most powerful GenAI video tool yet. So the next logical question becomes: “okay Mark, what content did Meta scrape to train its powerful new AI video model?” Here’s Meta’s very unsurprising response: “We’ve trained these models on a combination of licensed and publicly available data sets.” Yes, there are those three words again — “publicly available data.” And, by now, all of us know what that means. And what it doesn’t (i.e., consent and compensation).
(3) Watch These Copyright Tea Leaves Very Closely
A few weeks back a trio of judges on the U.S. Court of Appeals of the D.C. Circuit expressed continued skepticism that AI generated “creative works” (in this case an AI generated painting) are capable of copyright protection.
So how much human “input” is needed to preserve copyright protection (i.e., exclusive content monetization) in the context of GenAI — protection that is at the core of all creative endeavors (and the motivation to invest in them?).
No one knows. It’s all gray. But so far GenAI promoters haven’t succeeded in convincing either the courts or the Copyright Office that AI-only generated works are capable of copyright protection. So creators utilizing GenAI tools — and you studios and investors who fund them — always keep that in mind, since over-reliance on GenAI may lead to ill-conceived under-protection for your creative works.
(4) This Will Help You Sober Up From Your Wild Weekend
Today’s “fun f-AI-ct”: reportedly “57% of content on the internet today is either AI-generated or translated using an AI algorithm.” And that’s today. Imagine what it will be in the months and years to come. And as “synthetic” content builds and builds, many data scientists predict the real possibility of so-called “model collapse” (not a pretty picture, as reflected in this image below).
II. The Showc-AI-se
(1) GenAI Video of the Week
Award-winning GenAI filmmaker Paul Trillo, whose trailblazing works I’ve featured before, showcases one of his latest films — titled “A Love Letter To …” — which reveals not only the state of GenAI video today, but also its power and possibilities for marketing and advertising. Just press play to watch it. You’ll see what I mean ….
(2) GenAI Podcast of the Week
This week, my “ch-AI-rming" synthetic co-hosts discuss my recent article, “AI & Infringement: How the Supreme Court Will Decide (& What It Means for Media, Entertainment & Creators”). I generated the podcast simply by dropping my article’s link into Google’s new NotebookLM. No prompts of any kind needed. Insane.
The podcast is remarkably insightful and entertaining — I urge you to listen to my “co-hosts.” The podcast both accurately captures the relevant copyright issues and demonstrates the power of this new tech that’s so good, it’s a bit scary.
(COPYRIGHT NOTE: To be clear, I’m not against the use of GenAI. I embrace new tech, so long as it’s done right. GenAI developers scraping copyrighted works to train their AI models — without consent from, and compensation to, rights-holders — is not “doing it right.” It’s outright taking. I believe the U.S. Supreme Court will ultimately side with copyright owners and reject Big Tech’s “fair use” arguments.)
(3) GenAI Event of the Week
Join me this Friday night, October 11th, 5-8 pm, at this exclusive Google sponsored GAIN LA Innovator’s Night event at The Preserve LA in Hollywood (register here - space extremely limited), which is part of Andreessen Horowitz’s LA Tech Week. I’m honored to have been asked to moderate an all-star panel that features leading GenAI filmmaker Paul Trillo (whose short film is featured above), together with Google’s Jennifer Haran, and AI LA’s Todd Terrazas. (See COPYRIGHT NOTE above).
III. The mAIn event - Governor Newsom Walks the Tightrope to Create GenAI Law Where No Man (or Woman) Has Gone Before
It’s not easy being Governor of the Great State of California — and the generative AI “space” is no exception. California is “ground zero” for “The Great GenAI Copyright Debate” that frequently fills my newsletters. SoCal’s Hollywood and creative “right brain” heads butt up directly against NorCal’s Silicon Valley’s analytical and engineering-focused “left brain” heads. Both sets of heads — housed in their respective bodies of entertainment innovation and tech innovation — find their homes right here in the Golden State amidst the sunshine. And somehow, Governor Newsom needs to represent them all.
And here’s the wild part. Over just this past month, Newsom has signed a whopping 17 bills covering the use of GenAI. That’s right, 17! Not all pit Big Media against Big Tech, of course. But several do. And that incredible tension — and virtually impossible balancing act of promoting both responsible AI and AI innovation — revealed themselves during the past weeks in an endless stream of AI-focused legislation that found their way to Newsom’s desk. Eyes around the globe watched to see how he’d deal with these bellwether bills, since the rest of the world follows California’s lead both in creativity and tech. After all, 32 of the world’s 50 leading GenAI companies find their homes on his political turf, here in California.
The result? Newsom gave “big media” and the creative community a patchwork of big wins by signing several closely watched bills, while he gave “big tech” its own major win when he went the other way on another. Not surprising. A split decision in a sense. Yet that result pleased neither side — also not surprising, since ties rarely please anyone.
First, let’s take some of the big “wins” for the creative community. Newsom signed AB 2013 into law, which mandates transparency from the major AI developers — requiring them to publicly disclose the datasets they use to train their GenAI models. Relatedly, he signed SB 942, which will require the major GenAI systems to both include provenance disclosures in generated results and make tools available to identify AI generated content.
Together, this new legislation creates yet another powerful incentive (perhaps one of the most potent ones) for GenAI developers to license the content they need to train their AI models (rather than simply take it). You’d think that rational GenAI actors wouldn’t want to be seen as being on the wrong side of the law (infringement v. fair use) and trying to hide the ball, especially when those Big Tech players (like Meta highlighted above) won’t be able to rely upon those infamous three words anymore - sing it with me: “publicly available data.” They apparently will need to cough up the goods and go clean!
Are the “fair use” dominoes starting to fall? [NOTE: That’s next week’s newsletter — I’ll stack up those arguments … and then knock them down one by one — so stay tuned for that. I may convince you doubters … you know who you are!]
But wait. There’s more. Inspired by last year’s SAG strike, Newsom also signed AB 2602, which will allow actors to back out of existing contracts if studios use intentionally vague language to justify their use of digital clones of the performer’s voice or likeness without new consent and compensation. Newsom also signed AB 1836, prohibiting the creation and distribution of digital clones of a deceased person’s voice or likeness without prior consent from their estate; and SB 926 which protects all individuals (not just celebrities) from digital sexual harassment (e.g., the kind of deepfake sexual abuse Taylor Swift suffered earlier this year).
How will California enforce these new GenAI laws? No one really knows. But it’s a start. A significant one.
So what was Silicon Valley’s big win? Newsom vetoed SB 1047, the "Safe and Secure Innovation for Frontier Artificial Intelligence Models Act” — which would have required the major GenAI developers (OpenAI, Microsoft, Google, etc.) to implement so-called AI safety measures, including “kill switches” to prevent “critical harms” to the public and other dire consequences of AI run amuck (arguably, that “muck” could include some kind of creative calamity).
Newsom justified his veto of that bill — the only one that truly addresses regulation of the underlying AI models themselves (rather than specific uses of those models) — by saying that it focused too narrowly on large generalized AI models, thereby giving a false sense of security and overlooking smaller “fine tuning” models. In other words, he punted — seemingly saying that it was too draconian, yet not broad enough? Huh?
In any event, the end result of this wild GenAI legal barrage is a patchwork of California legislation designed to address specific identified harms. And that’s the problem — it’s a reactive approach (not a critique, just reality), rather than a systemic and overarching federal approach that proactively addresses potential broader abuse enabled by the underlying AI models themselves. Even leading GenAI scientists believe that such a broader approach — concededly brutally difficult to effectively define — is needed to address potentially catastrophic consequences that we can’t even anticipate.
How can reactive piecemeal legislation ever keep up, especially if done on a state by state (rather than federal) level? Patchworks — across 50 states — lead to uncertainty and “friction.” And that slows down GenAI adoption. So, while Big Tech may cheer Newsom’s big veto of SB 1047 now, the end result ironically may be counter to its interests. Friction causes the entire Hollywood and creative community to think twice before speedy GenAI adoption, due to the legal and relationship risks involved.
Nonetheless, with no federal regulation in sight (shocking, I know!), Newsom and California should be applauded for leading the way to address GenAI’s myriad risks.
And California’s approach ultimately may become the de facto standard for the country as a whole. (NOTE: let’s not forget that it’s a big world out there with over 200 countries, all of which have access to the same GenAI models — but under very different legal rules of the g-AI-me).
IV. AI Litigation Case Tracker - Updates on Key AI Litigation (now brought to you by McKool Smith)
Partner Avery Williams and the team at McKool Smith (just named “Plaintiff IP Firm of the Year”) lay out the facts - and latest critical developments - via this link to the “AI Litigation Tracker”. You’ll get everything you need (including my own separate detailed analyses) for each of the cases listed below. So much you need to know. So little time. We do the work so you don’t need to!
Here’s the central theme in all these cases: GenAI developers need access to the world of content for training their AI models (it’s all just “data” to them). But copyright owners understandably believe (as do I) that those tech companies must seek consent from — and pay compensation to — the relevant content/copyright owners whose content they scrape. And when they don’t, they increasingly find themselves in court.
(1) The New York Times v. Microsoft & OpenAI
(2) Sarah Silverman v. OpenAI (class action)
(3) Sarah Silverman, et al. v. Meta (class action)
(4) UMG Recordings v. Suno
(5) UMG Recordings v. Uncharted Labs (d/b/a Udio)
(6) Getty Images v. Stability AI and Midjourney
(7) Universal Music Group, et al. v. Anthropic
(8) Sarah Anderson v. Stability AI
(9) Authors Guild et al. v. OpenAI
NOTE: Go to the “AI Litigation Tracker” tab at the top of “the brAIn” website for the full discussions and analyses of these and other key generative AI/media litigations. And reach out to me, Peter Csathy (peter@creativemedia.biz), if you would like to be connected to McKool Smith) to explore these and other legal and litigation issues. I’ll make the introduction
And check out my firm Creative Media and our AI-focused services
Send feedback to my newsletter - including guest essay submissions and other ideas - to me, Peter Csathy, at peter@creativemedia.biz.