Getty Images Sues AI Image Generator for Copyright Violations
The intersection of artificial intelligence and copyright law is still being mapped
In the Future, All Yearbook Pictures Will Be Flattering
Given all the attention surrounding the release of OpenAI’s ChatGPT, it might have been easy to miss a recent lawsuit filed by Getty Images in Delaware Federal Court regarding Stable Diffusion.
Stable Diffusion is a cousin of ChatGPT—or, more properly, of ChatGPT’s sibling DALL-E (okay, okay, if it’s ChatGPT’s cousin, it has to be DALL-E’s cousin, too; it’s a metaphor, jack!)
In its most familiar form, Stable Diffusion, like DALL-E, generates strikingly sophisticated images based on human text descriptions; like OpenAI’s offering, it has been “trained” on the differences between billions of discrete images to have the eerie ability to create images that match even outlandish requests. Check out some examples here.
Remarkable, isn’t it?
Copywrong?
In the media hoopla surrounding the release of these amazing tools, their copyright implications have mostly flown under the radar.
But, inevitably, copyright cases against AI tools are beginning to be filed here and abroad by stock image behemoth Getty Images and will need to be watched. The latest case, filed on February 3, accuses Stability AI (the creator of Stable Diffusion) of copyright and trademark violations under the 1976 Copyright Act, the Lanham Act and Delaware unfair competition law.
“Stability AI has copied at least 12 million copyrighted images from Getty Images’ websites, along with associated text and metadata, in order to train its Stable Diffusion model,” the complaint states.
But that’s not all. Getty alleges Stability AI knew it was “scraping” copyrighted content to train Diffusion, and, in the process, sometimes removed copyright notices or spat out images that contained “a modified version of a Getty Images watermark, creating confusion as to the source of the images and falsely implying an association with Getty Images.”
Getty is seeking, among other relief, an order destroying “all versions of Stable Diffusion trained using Getty Images’ content without permission” as well as profits attributable to the infringement.
The Takeaway
So the stakes are high. And even if the legal implications aren’t yet clear, the questions that are being raised are ominous—and enormous. If ChatGPT, for instance, is trained on text gathered from the Internet, is any of what it produces susceptible to infringement claims based on copyrights in the underlying works used to train the models? Can a normal copyright infringement analysis be applied? Is scraping text for the purposes of training these models a fair use? Are the generated images and content produced by DALL-E, Stable Diffusion, ChatGPT and similar programs eligible for protection, and if so, who owns those rights?
Keep your eyes and ears open. And before you start using any of these AI tools to generate content, make sure that you’re okay living in the grey area for now.
COVID Prevention Claim Turns Out to Be...TRUE?!
Air purifier company told the truth about removing viruses from the atmosphere
The Prolonged COVID Era
Every time we start to get our hopes up that the “COVID era” is drawing to a close, something happens that seems to prolong the pain. A new variant pops up, a new wave starts just when people begin to return to “normal,” or winter rolls around and everyone goes indoors. Or all three at once.
But what hasn’t changed is the heavy regulatory focus on COVID claims in the advertising space, as evidenced by a recent National Advertising Division decision.
COVID-Generalization
The case concerns ACCO Brands USA, maker of the TruSens Air Purifier, which it claimed is “effective at eliminating 99.9% of the airborne coronavirus or COVID-19.”
ACCO had qualified this claim, adding, “Results from independent third-party testing using aerosolized airborne concentration of human coronavirus 229E over a two-hour period in a sealed chamber. HCoV-299e is a well-established surrogate for SARS-CoV-2.” NAD started poking around, and ACCO provided the scientific receipts.
As NAD proceeded, ACCO changed the qualifier again. The new text read: “Results from independent third-party testing using HCoV-229E, a human coronavirus with similar shape and size to COVID-19, which was aerosolized over a two-hour period in a sealed chamber. HCoV-229E is a well-established surrogate for COVID-19.”
And what did NAD find?
The Takeaway
“NAD determined that ACCO Brands’ testing provided a reasonable basis for the challenged claim...[and] also found that ACCO Brands products effectively filter SARS-Cov-2, providing additional support for the advertiser’s claim that “TruSens Air Purifier is effective at eliminating 99.99% of airborne coronavirus or COVID-19.”
What?!
A COVID claim turned out to be true.
We’ve covered so much COVID chicanery over the past three years, so much fraud and deception, that finding out that a company is actually telling the truth about a COVID-19 prevention claim is putting us in strange psychological territory.
But when we take a step back, it’s clear this result confirms two things we already knew: (1) COVID and other health claims are always going to attract scrutiny and (2) although such claims are not per se unlawful by any means, you’d better have the receipts.
FTC Pulls Trigger on Health Data Rule
GoodRx is called out for the wrong type of data transparency, fined $1.5 million
Do You Feel Lucky?
The great Russian playwright Anton Chekhov is responsible for the famous maxim “One must never place a loaded rifle on the stage if it isn’t going to go off. It’s wrong to make promises you don’t mean to keep.”
This is excellent advice for the dramatist, but what does it have to do with ad law?
Well, way back in act one—we’ll call it 2009—the Federal Trade Commission introduced its Health Breach Notification Rule. The rule, which was revised in 2020, requires “vendors of personal health records” and their service providers “to notify consumers following a breach involving unsecured information.”
And there the rule hung, loaded, above the metaphorical fireplace, until last month.
“Transparent” Is in the Eye of the Regulator…
In February of 2023, the FTC launched its first-ever HBNR complaint against GoodRx Holdings, an internet health platform that offers drug discounts and other health services and boasts that it hopes to “make prescription pricing and healthcare more transparent.”
An unfortunate choice of words, perhaps, given the Commission’s accusations.
“GoodRx allegedly violated the FTC Act by sharing sensitive personal health information for years with advertising companies and platforms—contrary to its privacy promises—and failed to report these unauthorized disclosures as required by the Health Breach Notification Rule,” the Commission states.
This behavior included several nasty allegations under the FTC Act: That GoodRx promised its users that it would “never share personal health information with advertisers or other third parties” but nonetheless spilled users’ prescriptions, health conditions and other personal information to ad platforms. That it used that info to create targeted ads aimed at its own users. That it failed to put constraints on the use of that data by its third-party vendors. And that it failed to establish “sufficient formal, written, or standard privacy or data sharing policies or compliance programs.”
The Takeaway
So there’s the underlying alleged bad behavior, and then there’s the failure to inform consumers. The latter is a first-of-its-kind action, and it’s going to cost GoodRx a pretty penny.
Under the order that settled the complaint, GoodRx will fork over $1.5 million in civil penalties and be prohibited from sharing health data for advertising purposes. Additionally, the company will need to secure affirmative express consent before disclosing data for any reason and subject itself to a data retention schedule (among other things). Bang!
The Health Breach Notification Rule sprung out of the oughts, when personal data was just beginning to be knit together, and the quaint language of its original announcement makes it clear: “Consumers can connect a device such as a pedometer to their computers and upload miles traveled, heart rate, and other data into their personal health records,” the Commission stated. “These innovations have the potential to provide numerous benefits for consumers, which can only be realized if they have confidence that the security and confidentiality of their health information will be maintained.”
You don’t say? Well, for better or worse, we’ve come a long way since 2009. In fact, it’s amazing that it’s taken this long for the Rule to be deployed; but the curtain is just beginning to rise on big data’s second act.
’80s Pop Icon Rick Astley Sues ’20s Rapper Yung Gravy
But it’s not for sampling…it’s for that sweet, sweet voice
Rickbait
This is a twisted one.
First of all, we all know or remember this guy, right?
(See, this is a story about Rick Astley, so we neutralized the Rickroll right out of the gate. We told you to click on a link, because we wanted you to see Rick Astley in all his glory. We’re serious about ad law, and we’re not going to indulge in dated pranks. So go ahead—bask in Rick, unapologetically. You weren’t tricked into seeing him; you’re right where you want to be.)
But! You might not remember (or know) this guy. In fact, we hope you don’t know him. Really—don’t click on that link. Don’t Do It.
The upstart in the second link is a rapper of dubious talent named Yung Gravy. But if you ignored our advice and did, in fact, click on the link to Yung’s video, you’ve probably got a decent idea of where this story is headed.
Ricklayers
It’s about Rick Astley’s voice, of course. His smooth-as-butter-left-out-on-the-countertop baritone, featured most memorably in his 1987 solo hit “Never Gonna Give You Up.” This mega hit from days of yore is clearly sampled in Yung’s…opus, we guess we’ll call it…called “(Betty) Get Money.” It seems likely that Rick is suing him for copyright infringement, right?
Wrong! Rick is suing Yung for right-of-publicity violations (the complaint was filed in Los Angeles County Superior Court at the end of January). How do we make sense of this charge?
Well, Yung and his co-defendants didn’t sample Rick’s voice from “Never Gonna Give You Up.” They licensed the right to use the melody and lyrics in the song from a third party; Rick didn’t write the song, so he couldn’t block licensing of the underlying composition.
So while Yung couldn’t sample Rick’s original recording—he wanted to, but Rick blocked his efforts—he could hire this dude to sing it just like Rick.
Anyone less smooth and resourceful than Rick might have thrown in the towel at this point, because copyright law doesn’t prevent someone from making a “soundalike recording.” Rick didn’t really have grounds to sue on that score.
So, instead, he pursued right-of-publicity claims as well as false endorsement and false designation of origin under the Lanham Act.
The Takeaway
It’s not as crazy as it sounds; the complaint is anchored in a case brought by Bette Midler, of all people, back in 1984. In that case, the divine Miss M prevailed over a major motor company, which acquired the rights to the composition underlying one of her songs and hired an impersonator to sing it.
The only problem for Rick is that Midler’s imitator was featured in an advertisement, not an expressive work. Say what you want about Yung’s track, but it was only selling him. Yung may be protected by the First Amendment.
This is shaping up to be a fascinating case, as it will address whether or not the right of publicity can make the leap over to cover artistic endeavors. For an excellent—and very funny—discussion of the Rick and Bette cases, check out CopyrightLately’s article on the pair.
And to cap off this wild ride, we note with some glee that legions of Yung Gravy fans have begun using “(Betty) Get Money” to “Gravyroll” unsuspecting web-surfers. Check some of their reaction videos out here.
Logan Paul Sued Over Animal NFT Rugpull
If you didn’t understand that headline, read on
Logang
We love influencers. We really do. The best among them embody an entrepreneurial spirit and creativity that any marketing executive would envy, and they’re out there doing it all alone (mostly). They make us feel...protective.
The worst among them, however…
Consider the case of Logan Paul. After ascending to the heights of the now-defunct Vine app, Paul swung over to YouTube, where he amassed a ridiculous number of subscribers to his channels and his videos generated billions of views. How did he amass the following he calls the “logang”? Controversy after controversy.
Crass Menagerie
But the latest scandal he’s attached his mug to threatens to deliver more than just deleted videos and obligatory apologies. In 2021, Paul launched and promoted (for a time, at least) CryptoZoo, a crypto-based game that generates non-fungible tokens of strange hybrid animals and promises cash rewards for player participation. The project’s white paper claimed CryptoZoo was developed to “introduce blockchain technology, specifically NFTs, to new market participants by combining the proven success of gamified animal ownership with the anticipatory optimism of collecting trading cards.”
Which may be the most 2021 sentence ever written.
The game itself is too baroque to explain here, but the basics are this: Users spent cash to buy crypto tokens, called “zoo.”. These coins would be used to purchase egg NFTs, which would hatch various animals. The user then could mate the animals to create weird hybrids, which would earn them zoo tokens. Those tokens could, conceivably, be cashed out, or used to buy more “zoo.”
Controversy-addicted influencer? Cryptocurrency? Passive income?
Cue lawsuit in 3...2...1…
Zoo Suit
Logan’s venture hatched a class action against the game company CryptoZoo, Inc., Paul, and a handful of co-defendants in Texas’ Western District. Lead plaintiff Don Holland, a police officer and crypto investor from Round Rock, bought into the game in 2021. He claims that he sunk his money into the game because he “thought there was long-term merit in the project due to the false affirmations by the development team.” A “steady drip of information on positive developments—which were untrue” led him to invest more cash in the game, reaching a total of $3,000 sunk.
“Once he saw Logan Paul announce another NFT project and started [sic] ignoring ZOO,” the complaint states, “he realized his money was likely lost. CryptoZoo was never released as advertised and the value of the Zoo Tokens and...NFTs plummeted.”
Holland claims the entire enterprise was a “rugpull”—cryptobro speak for fronting a project that doesn’t exist to make profit on uninformed investors. Naive cash flows in, and the producers of the platform stuff their pockets and then shutter the business and head out the door.
As Holland explained: “June 11, 2021, was considered internally by Defendants as ‘Zoo Day,’ the day upon which they released—without any public notice—their digital products for purchase on the Binance blockchain. On Zoo Day, and until the release was publicly announced, Logan Paul [and co-defendants] purchased these digital products at an artificially low value. Soon after the project was publicly announced, [defendants] sold large amounts of the digital products for an immediate and large profit, effectively stealing the money of consumers who had invested.”
The Takeaway
The CryptoZoo meltdown, which is quite small in comparison to more recent crypto-related and -adjacent scandals like the FTX debacle, is something of a saga.
There’s the dramatic intervention of CoffeeZilla, a YouTube personality who posted a withering three-part takedown of CryptoZoo in mid-January, before the case was filed. And the ensuing war of words between CoffeeZilla and Paul, which unsurprisingly ended with a public apology from Paul and the release of his “3-step plan for CryptoZoo, including a $1.3M rewards program for disappointed players.”
Holland’s suit itself is brought by yet another YouTuber, called Attorney Tom.
The lesson here is that when it comes to crypto, the entities keep multiplying, and they all seem to be social-media-related. It’s as if social media controversy, for the time being, is built into the DNA of ventures that are tied to blockchain.
Be careful out there. Not as an investor—we’re not here to give that kind of advice. As a businessperson. Avoid alliances with influencers who court controversy; ignore the magnetic pull of millions of followers and billions of views. Make sure the underlying endeavor is sound, because for a while yet, crypto-derived marketplaces are a forest you don’t want to wander into.
[View source.]
