Brands Using Sora: Lawsuits 'R' Us?
Legal Uncertainty Slows Generative AI Adoption (& Begs for a Comprehensive Licensing Solution)
Welcome to your Monday brAIn dump! First, “the AI:10” - the 10 key AI-focused headlines for media and entertainment last week. Next, this week’s “mAIn event” (my “take” about whether brands like Toys ‘R’ Us open themselves up to copyright infringement when they use OpenAI’s Sora and other genAI tools). Then, the “AI legal case tracker” - updates on the key genAI copyright infringement cases (expanded this week to include two new blockbuster cases against generative AI music platforms Suno and Udio). Then a final toast with “the cocktAIl”: AI mixology.
[NOTE: no newsletter next Monday due to the holiday - am resting my brAIn for the week.]
You can now listen to “the brAIn” PODCAST!
For those of you who’d rather listen than read, we got you! Check out this week’s episode of my new companion podcast via this link (or find “the brAIn” on your favorite podcast network - click on buttons below to find on Apple or Spotify). And please be sure to follow it, rate it, and share it.
I. the AI:10 - last week’s 10 key AI headlines
(1) Generative AI music companies face the music! What a week, as the RIAA filed twins-ie litigations on behalf of the major record labels against Suno and Udio, two of the most well funded AI music generators. Not surprising - I had listed both of them on my “Top 10 ‘Watch List’” just a few weeks ago. Both scrape “publicly available data” - those three notorious words that almost certainly mean copyrighted works (but they’ll never tell!). Read the RIAA’s press release. I’ve added both to my “AI Case Tracker.”
(2) As if that weren’t bad enough, Suno is being used to generate racist and antisemitic music. Yup, it’s true, according to a new report from the Anti-Defamation League. Read more here via Music Business Worldwide.
(3) Meanwhile, YouTube is trying to convince the major labels to license their music for its AI song generator. Good luck with that! Reportedly, YouTube is showering the labels with boatloads of cash to convince them to sing its AI tune. But they’re not convinced yet. Read more here via Ars Technica.
(4) All of this begs the question, should media companies litigate or license? Here’s a good scorecard about it all via TheWrap. And here’s one more to add to the list. Non-profit newsroom The Center of Investigative Reporting (producer of Mother Jones) just sued OpenAI for infringement, essentially using a “market substitution” argument. Read more here via The Hollywood Reporter.
(5) Wanna make your own TV show? Now you can! Fable Studio’s “Showrunner” is the new generative AI platform that reportedly makes it all possible. Users can create original episodes of their favorite animated television shows. But did Fable Studio get consent from the IP owners to do it? Read more here via TheWrap.
(6) Can any of us tell what’s real now? AI is getting better fast, reports The New York Times. Now it’s getting close to being virtually impossible to distinguish whether photorealistic images are, in fact, real. You can see for yourself via this link. But at the same time, Instagram's new “Made with AI” tag is enraging photographers for flagging photos that are NOT AI-generated. Ironic, don’t you think? Read more here.
(7) It’s farming news, AI style! AI newscasters in India deliver news that moos. Is the news they deliver synthetic too? Read more here via Bloomberg.
(8) Maybe that’s why the number of dystopian movies have doubled since 2010. And I certainly don’t expect any slowdown anytime soon. Read more via TheWrap.
(9) Jeff Bezos demands Amazon-backed Anthropic make faster deliveries! And it listened. Last week, Anthropic announced its new souped up genAI model “Claude 3.5 Sonnet” to catch up to ChatGPT-4o. But expect OpenAI to leapfrog over Bezos later this year when it is expected to release GPT-5. Read more here via Wired.
(10) Nvidia doesn’t just serve us AI chips anymore. It recently announced its new NVIDIA ACE generative micro-services dip to accelerate the next wave of digital humans, further blurring the lines of man and machine. Read more about it here from Nvidia itself.
II. “the mAIn event” — Brands Using OpenAI’s Sora: Lawsuits ‘R’ Us? (the risk of litigation slows adoption and demands a licensing solution)
Last week Toys ‘R’ Us – the once lofty and ubiquitous child’s brand – capped its reunion tour with ad execs at Cannes Lions with the premiere of a new 66-second ad it hoped would signal to the world, “we’re back!” Its new toy that took center stage was no elf on a shelf. It was OpenAI’s controversial AI video generator Sora, and Toys ‘R’ Us’s headline was that it was the first major brand to produce a major spot using Sora and its undeniable power. Reviews of the ad itself were decidedly mixed, but the ad’s real impact may not have been at Cannes — but rather in the halls of enterprising plaintiff’s law firms.
Toys ‘R’ Us’s use of Sora flags a potential expansive new front in the generative AI copyright wars that are fast proliferating in the courts. And brands may soon find themselves swept up into that front. That threat alone slows down overall big brand adoption of Sora and begs for a comprehensive copyright licensing solution.
OpenAI Is At the Center of the Copyright Storm
OpenAI, of course, finds itself at the center of AI’s copyright infringement storm. Media companies like The New York Times and individual creators like comedian Sarah Silverman argue that OpenAI’s unlicensed scraping of their copyrighted works — in order to build billions (if not trillions) of market substitution for itself — constitutes blatant infringement. For its part, OpenAI – and virtually all other tech companies accelerating the generative AI revolution – cry “all’s ‘fair use’ in love and war.” It’s all one big misunderstanding. That’s the same refrain two of AI’s most celebrated music generators — Suno and Udio – sang last week too as they brawled with the RIAA (which had just taken them to court, as indicated in the AI:10 above).
If the courts do ultimately come down on the side of media and the creative community on the input side of the question (i.e., the unlicensed scraping of copyrighted works – which I believe they will using a similar “commercial harm” rationale identified by the U.S. Supreme Court in the Andy Warhol case (you can read my detailed analysis here) – then a credible argument can be made that big brands like Toys ‘R’ Us and others that use Sora are liable for infringement too.
Toys ‘R’ Us: Could It Be Secondarily Liable for Using Sora?
Companies can be secondarily liable for acts of copyright infringement if they: (1) directly benefitted financially from the infringing activity; (2) had the right and ability to supervise or control that activity; and (3) failed to exercise that right and ability. That’s not me talking. That’s the 9th Circuit Court of Appeals’ 3-part jury instruction on the subject. And that’s arguably the case with Toys ‘R’ Us and Sora.
Toys ‘R’ Us clearly benefitted financially from the arguably infringing activity of using Sora to create its ad. You can find the global press talking about it everywhere (just Google it). Element 2 of the relevant test is also satisfied — Toys ‘R’ Us had the ability to supervise or control the making and showcasing of its ad. And then, to fulfill the 3-part test trifecta, it failed to exercise that right and ability.
To be clear, we don’t know how the courts ultimately will decide the fundamental copyright infringement issues facing OpenAI and other Big Tech AI generators. But it’s hard to argue that big brands like Toys ‘R’ Us aren’t well aware of the current AI copyright infringement controversy surrounding AI’s mass scraping of unlicensed works. And merely because there’s no certainty either way doesn’t necessarily absolve the user of allegedly infringing products from secondary liability, especially when the fundamental issues at stake are so open and notorious.
So if the courts ultimately find that OpenAI is infringing at a massive scale, those rulings may sweep Toys ‘R’ Us and other big brand users of Sora and generative AI tools into the liability morass. Again, not saying copyright owners would win. But plaintiff lawyers can make that case. Credibly. And that threat alone slows overall big brand — and big media — adoption of Sora and all other generative AI platforms.
OpenAI Would Indemnify Toys ‘R’ Us
To be clear, it’s virtually certain that OpenAI indemnifies users like Toys ‘R’ Us (and its agency) for any damages caused by their use of Sora. But indemnification isn’t the same thing as absolution (or legal risk mitigation). A finding of infringement is still infringement, and this more than theoretical risk is enough for many big brands to pause before putting their pedal to the AI metal. I speak with “insiders” all the time about these things, and they’ve confirmed that many brands, agencies, and media companies are moving very cautiously at this point while the copyright jury is still out.
That’s a smart strategy given what’s at stake, including inevitable blowback from the creative community itself for having no “say” in the AI training decision – and no commercial reward for the value IP owners give the generative AI companies in the first place. Let’s not forget that this revolutionary new technology is useless without the content that feeds and fuels it.
All this legal uncertainty and undefinable risk dampens the appetite to go full bore on using these generative AI tools right now. So it’s in OpenAI’s and Big Tech’s interests to solve these fundamental copyright licensing issues fairly and fast — with new comprehensive licensing schemes that recognize IP’s transformational value.
Send me your thoughts to peter@creativemedia.biz.
III. the AI legal case tracker - updates on key AI litigation
I lay out the facts - and the latest developments - via this link to the “AI case tracker” tab on “the brAIn” website. You’ll get everything you need (including my detailed analysis of each case) of the cases listed below. NOTE: TWO IMPORTANT NEW CASES ADDED THIS WEEK TO THE TRACKER!
(1) UMG Recordings v. Suno
(2) UMG Recordings v. Uncharted Labs (d/b/a Udio)
(3) Universal Music Group, et al. v. Anthropic
(4) The New York Times v. Microsoft & OpenAI
(5) Sarah Silverman, et al. v. Meta
(6) Sarah Silverman v. OpenAI
(7) Getty Images v. Stability AI and Midjourney
NOTE: Go to the “AI case tracker” tab at the top of the page at “the brAIn” website.
IV. the cocktAIl — my mixology of “must attend” AI events
After all, it’s always happy hour somewhere!
COMING SOON - LESS THAN ONE MONTH AWAY IN JULY! It’s Digital Hollywood’s first major comprehensive generative AI-focused conference, “The Digital Hollywood AI Summer Summit”, taking place July 22nd - 24th. It’s all virtual and entirely FREE. Learn more here via this link. So many sessions, so many great speakers. I moderate two great panels filled with superstars.
Check out my firm Creative Media to learn more about the unique power, reach and impact of our business consulting and development services, together with Peter’s corporate speaking and AI business workshops (both virtual and IRL).
Send feedback to bizdev@creativemedia.biz.