Arts advocate. Warning about libertarian definitions of freedom that fail to consider asymmetrical power since 1996. Secular humanist writing about religion.
Let’s be really clear here. The AI companies that don’t want to negotiate with artists over usage terms for training AI have sought to portray this as a fight over technology. As the past resisting the future. It is not. Artists are contesting exploitation, not technology.
1/__
“Andreessen Horowitz is warning that billions of dollars in AI investments could be worth a lot less if companies developing the technology are forced to pay for the copyrighted data that makes it work.”
This is NOT from the
@TheOnion
“If a federal judge finds that OpenAI illegally copied The Times' articles to train its AI model, the court could order the company to destroy ChatGPT's dataset, forcing the company to recreate it using only work that it is authorized to use.”
“Today lawmakers from both sides of the aisle agreed that OpenAI & others should pay media outlets for using their work in AI projects. “It’s not only morally right,” said
@SenBlumenthal
“It’s legally required.”
—
@Knibbs
“I’ve resigned from my role leading the Audio team at Stability AI, because I don’t agree with the company’s opinion that training generative AI models on copyrighted works is ‘fair use’.”
🙌
A number of people have asked me why I get so bothered by crypto/blockchain/NFTS/Web3 when no one is making me participate against my will. It has come up enough times that I thought it merited a response. A thread.
“The VC firm said AI investments are so huge that any new rules around the content used to train models ‘will significantly DISRUPT’ the investment community's plans and expectations around the technology.”
This from the folks that only ever use “disruption” as a good thing. 🤦🏻
NEWS ALERT…
Class action lawsuit filed against Midjourney, StabilityAI (StableDiffusion) & DeviantArt for the unauthorized use of copyright protected works to train their AI systems.
“The promise of AI is that it will generate the image in your imagination if you can describe it. Yet this promise is limited by whether or not someone else has already created part of the image you imagined.”—
@itsjamestapper
🙏🏼 AI can’t imagine anything not in its training set.
"I’ve started to worry that all of this stuff—Web3, NFT's, even the blockchain—is just a Ponzi scheme to cheat people like me out of our money, & no one even thought to explain this to me outside of the hundreds of people I deliberately ignored."
@TheOnion
Hey
@DaveJWVillalva
. Were all these copyright protected characters properly licensed? If not, you’ve just confessed to massive copyright infringement.
120 days ago, I’d never generated an AI video.
Now I make them every day.
These Super Friends & Foes were among the first I created.
They got 10k+ views, not bad for 500 followers.
Here’s why it worked and how I made it 🧵
Anthropic, like OpenAI before it, makes the unintentional—but compelling—case for banning general purpose AI models:
“If licenses were required to train LLMs on copyrighted content, today’s general-purpose AI tools simply could not exist.”
China’s draft AI regulation: “providers of generative AI must ensure data used for training & optimization is obtained through legal means, & such data must…not contain content that infringes intellectual property.”
While I may disagree with other provisions, China gets this 🎯
@daveweigel
@nilerodgers
“The Disco Sucks movement, where they held big events and burned our records, was racism. It wasn’t about them not liking the music; they were scared of the societal change that disco was bringing about.”
👇🏽💙
“If we see in AI, in light of its meaningless sociopathy, something that “learns just like humans do” & so “makes art just like humans do” we should be ashamed of our low opinion of humanity, of the smallness of our imaginations, & of the gruel we are willing to settle for.”
Even if they solve the theft, I don’t care to see or read or listen to what the algorithm extrudes. I want my art to come from real humans with lived experiences and points of view. I never knew some others didn’t care, until AI exposed the rift…imo, art must be human to matter.
On the topic of
#AIart
: a common defense for AI is that it learns from other art in the same way humans do, therefore it's not theft.
So I wanted to try to illustrate how that's not accurate.
I’m sorry, but “it” isn’t capable of being CREATIVE. AI models lack any form of intentionality. They can be used to produce unique images/text (while also being wholly derivative), but to call this an exercise in creativity is to diminish the very notion of what creativity means.
Okay, let’s talk exploitation. Did the artists, musicians, songwriters, visual artists etc whose creations were used to train the AI models all consent to the use of their works? If not, why are you advertising them? Is the end of the world not coming quickly enough?
“Today’s generative AI models can clearly be used to create works that compete with the copyrighted works they are trained on. So I don’t see how using copyrighted works to train generative AI models of this nature can be considered fair use.”
—
@ednewtonrex
“I’ve resigned from my role leading the Audio team at Stability AI, because I don’t agree with the company’s opinion that training generative AI models on copyrighted works is ‘fair use’.”
🙌
“OpenAI Says It's Fine to Vacuum Up Everyone's Content and Charge for It Without Paying Them”
That’s a headline without deceptive both-sideism. 🙏🏼
@futurism
@vtanger
“Without permission or payment, Defendants [OpenAI, Microsoft] copied Plaintiffs’ work to build a massive commercial enterprise that is now valued at billions of dollars.”
This is very interesting. In the end, the AI architecture isn’t so important. The gold in AI is the data. Some of us understood this intuitively. It seems natural that data refineries should pay for their principal resource, no?
interesting post from an OpenAI employee claiming that all large language models reach the same endpoint regardless of training strategy or clever tricks
this is of course what the bitter lesson teaches us but useful to get an up to date confirmation that it still holds true
“Nobody owes AI companies a right to exist, certainly not the owners of the content they have appropriated and mined for their own benefit.” 🙏🏼
@dominicyoung
The future of AI, and the future of the commercial viability of human creativity, comes down to a simple question: is it OK to use someone else’s work, without their permission, to train an AI?
by me for
@pressgazette
“If SD is able to generate new paintings ‘in the style of’ a living artist, that is likely to depress demand for all of that artist’s past & future work. And SD is only able to do this because it was trained on the artist’s previous work—without paying the artist a dime.”
whether I participate or not, these developments shape the world in which we all live. They’re best considered political movements, or manifestations of a political movement. That political movement is uber-libertarianism, & its goal is to destroy the institutions of democracy.
“We're all slightly buying into the con by even referring to all this as AI. These are unthinking algorithms that use the work of humans without consent, credit or compensation to regurgitate something similar but empty.”
—
@TomHumberstone
Artificial intelligence is an oxymoron.
The wonderful
@Aftermath_site
interviewed me the other day. I ended up discussing the impacts of AI a little: "Art made by humans will always exist but I think we're making it harder for people of all classes and backgrounds to make a living doing it."
Does AI “learn” like a human?
“It’s an entirely different category, with no shared characteristics. Analogies to the brain are just as misleading as when people used the same analogies to describe computers in the 1950s.”
—
@fchollet
🔥🔥🔥🔥🔥🔥🔥🔥🔥🔥
Deep learning takes data points and turns them into a query-able structure that enables retrieval and interpolation between the points.
You could think of it as a continuous generalization of database technology.
Defenders of current exploitative AI practices like to say that “real authors” shouldn’t worry since AI will never replace human genius. But they (intentionally?) fail to observe that flooding the zone with shit makes “real authors” invisible. Irrelevant.
#DrowningInIrrelevance
Hey artists—thought you might want to know that the only reason you think you should be paid for your work is because you were tricked by evil media companies into believing that your consent matters. Luckily, the Chamber of Progress (Google, Meta, Amazon et al) has your back. 🤯
What a depressing read. All these companies who like to talk about “AI for good” then argue that unauthorized use of creative works to build their products is morally fine & wholly justifiable. There’s no happy ending down this path—AI rooted in theft won’t produce “AI for good.”
I don’t wish to see this political movement succeed. It’s founded on the most anti-progressive idealization of freedom imaginable. A freedom in which any consideration of the consequences of one’s conduct is antithetical to liberty. A freedom that doesn’t consider relative power.
Tragic on so many levels. People generally fail to recognize how current genAI models—by using creative works without consent, effect a massive value transfer from a diverse creative community to a handful of Silicon Valley companies & VC’s.
@UNESCO
@WIPO
@raceip
@MarielzaTalks
“AI-generated Indigenous art is appearing on online marketplaces where art and derivative products are sold-often directly competing with the work from real Indigenous artists-despite…often having policies that are supposed to protect Indigenous culture.”
“Instagram will soon let anyone on the platform remix your new photos, as long as your account is public. There’ll be a way to turn remixing off, but you’ll have to actively opt out once the option is live — it’ll be enabled by default.” ☎️☎️☎️
@nppalawyer
@NPPA
@Unite4Copyright
“AI-art generators are trained on enormous datasets, containing millions upon millions of copyrighted images, harvested without their creator’s knowledge, let alone compensation or consent. This is effectively the greatest art heist in history.”
Interesting development on copyright & AI in France, reinforcing the importance of ensuring the digital economy creates opportunities for creators rather than undermining them. Consent and dignity must be our guides.
Thanks
@JusticeSila
!
The law would require AI software to obtain permission from authors or rights holders before using their works & any resulting creations must be labelled as "generated by AI" & include the names of the original authors.
#AIart
#AIcommunity
#AIArtwork
Great to see governments demanding the ethical/legal collection of data to train AI. Forget about a 6 month pause—we need an immediate & global shutdown of all AI models until they can demonstrate that all data is collected on the basis of consent.
Former GC of the Copyright Office (& perhaps the most knowledgeable copyright practitioner on earth), Jon Baumgarten, just released a letter to the House Judiciary Committee rebutting suggestions that training AI was protected by fair use. The analogy to photocopying is the 🧑🏽🍳💋
Web3 is the selling of a lie. A lie that depends on buy-in from those that can least afford it. A lie that brings casino mentality to all of society. A lie that serves as a form of regressive taxation in service of a deregulatory paradise for the wealthy.
At today’s Senate Judiciary hearing on Journalism & AI, Jeff Jarvis plans to testify that “the real question is whether AI should have the same rights that journalists & we all have: the right to read, the right to learn…”
Wait till he hears that AI can’t read or learn! 🤯
“I’m hard-pressed to understand how a system that rests almost entirely on the works of others, & can be commercialized or used to develop commercial products, owes nothing—not even notice—to the owners of the works it uses to power its system.”
—
@RepHankJohnson
#StripMiningArt
Altman says it’s “impossible” to operate ChatGPT unless there’s an exception allowing him to capture & use the world’s cultural output without consent to create his commercial product. “Impossible” may not mean what he thinks it means.
@LeightonAndrews
@IvorsAcademy
@crispinhunt
Rough Translation: We won’t get fabulously rich if you don’t let us steal, so please don’t make stealing a crime!
Don’t make us pay 𝘭𝘪𝘤𝘦𝘯𝘴𝘪𝘯𝘨 fees, either!
Sure Netflix might pay billions a year in licensing fees, but *we* shouldn’t have to!
More money for us, moar!
The “AI is just like a camera, capable of infringing & non-infringing uses” is probably the stupidest assertion I’ve ever heard. Does a camera come pre-loaded with all of the existing images in the world so that the user can generate derivative content paste?
@idonthaveaband
🤣"they want users to shoulder the legal responsibility of material generated by their systems".
(Are these the same people that promised to take on legal fees for lawsuits generated from using their Invasive models??) Lies...smoke & mirrors!
#FTC
Wow. I knew
@mer__edith
was amazing, but this is astonishing anyway. In a few minutes, she demystifies Silicon Valley’s VC culture & pierces the hype around GenAI. This is truly must-watch TV.
🔥🔥🔥🔥🔥/4
.
@mer__edith
tells
@cpassariello
, “VC's require hype to get a return on investment because they need an IPO or an acquisition … You don't get rich by the technology working, you get rich by people believing it works long enough that one of those two things gets you some money."
Another article that attacks a straw man, suggesting that artists are at war with AI rather than fighting against exploitation. Many artists may indeed not like the aesthetics of “AI art,” but that’s beside the point—the fight is against misappropriation.
This simply, but brilliantly, explains why forcing artists to “opt-out” from the use of their works to train AI is both unfair & unworkable. It is designed to fail. AI platforms must have a duty of care to ensure the integrity of their supply chains. No putting this on artists.
OpenAI now not only demands you tell them you opt out from their training data but also that you upload all your illustrations to their servers. How is this legal? This puts enormous costs on artists.
Quick math⬇️
A 🧵 for all you artists campaigning against unethical AI—keep it up. We are making progress. AI companies can no longer evade questions about the provenance of their data. Special shout out to
@kortizart
@SarahCAndersen
&
@Kelly_McKernan
.
1/…
AI is learning from stolen intellectual property. It needs to stop.
“This is wholly unacceptable behavior. Our books are copyrighted material, not free fodder for wealthy companies to use as they see fit, without permission or compensation.”
🎯🎯🎯🎯🎯
“Previous generations of creative tools weren’t built using copyrighted work. Users could infringe copyright with them, sure, by uploading or recreating copyrighted work themselves - but the companies who built the tools hadn’t used copyrighted content to build them.”
I hear lots of people argue that gen AI companies should be no more liable for their users’ copyright infringements than tools like Photoshop are. But this is totally false.
The difference with gen AI systems is that *they are built using copyrighted content*. They only work…
It’s really not that complicated. Artists & other creators should be paid for use of their works. They shouldn’t be indentured servants forced to cultivate the tools of their own demise. This isn’t about technology—it’s about human conduct. Let’s be the best version of ourselves.
@SashaMTL
“I was OK stealing the work of artists to train AI. I was OK paying moderation workers $2/hour to stare at horrible images. I was OK reifying bias & discrimination. I was OK with flawed surveillance tools being used to make important decisions. But I sense something bad coming.”
AI companies’ Defenses to Misappropriation Greatest Hits. It’s got it all—
copying is fair use
unauthorized use is necessary for Innovation
AI learns just like humans
AI extracts only non-protectable ideas, not expression
Creators wouldn’t get much $ anyway
& this is the 🧑🏽🍳’s 💋
One would hope that at some point—hopefully soon, we would stop needing to remind everyone that copies of protected works are reproduced without consent as part of the training process. OpenAI specifically noted this in comments filed with the US government.
BREAKING: OpenAI announced GPT-4 Turbo -- its most powerful AI model yet -- during its first in-person event Monday, and Microsoft CEO Satya Nadella made a surprise appearance. Plus lots more announcements in my roundup below
@CNBC
Artists will not give up until this issue is adequately resolved. AI companies are just delaying the inevitable. VC’s & others in the investment class may not care—they just want to get a return on their investments & don’t care that the parasitic sector will go up in flames.
Anyone that characterizes this as pro AI v anti AI artists is both failing to understand reality, & being a useful tool for exploitative AI companies. Any journalist that portrays this as the old guard resisting change is contributing to exploitation. So please don’t do it
2/__
“AI-art generators are trained on enormous datasets, containing millions upon millions of copyrighted images, harvested w/o their creator’s knowledge—let alone compensation or consent. This is effectively the greatest art heist in history.”
—
@art_inquiry
)
BREAKING: LAION recognizes that collecting & using “data” isn’t necessarily ethical. Will the next domino to fall be a rejection of “data” that’s been collected without consent? That would be monumental—a Revolution for the people, safeguarding the role of consent in a digital 🌏
big breaking news: LAION just removed its datasets, following a study from Stanford that found thousands of instances of suspected child sexual abuse material
If this isn’t a parody, I’m dead.
“I spent hours in Photoshop tweaking the images, creating my own text to make them look like a proper comic book. It was a labor of love, but I knew it would be worth it.”
“Labor of love!” That’s the 👩🏼🍳’s💋.
Interesting 🧵—especially:
“The steeper decline in earnings than jobs is particularly striking, because it means that not only is genAI directly taking away digital freelancers’ work, it’s also devaluing the work that they do still carry out.”
AI is a labor issue, 1st & foremost
NEW: Generative AI is already taking white collar jobs
An ingenious study by
@xianghui90
@oren_reshef
@Zhou_Yu_AI
looked at what happened on a huge online freelancing platform after ChatGPT launched last year.
The answer? Freelancers got fewer jobs, and earned much less
There's another website that advertises that they will make nude photos of anyone you want "no judgment." They also ask for you to be specific about what you want. Clearly, the images are AI generated, so my guess is that they copy-paste your request into a prompt box 1/2
Web3 is characterized as an evolution away from the centralization of Web2, but it is the same wolf in new clothes, & continues the Thiel/Bannon VC-enabled assault on our democracy. It enfeebles in the name of empowerment. Barlow on steroids.
“Instead of playing catch-up, we need to regulate generative AI at the source: its training data. We need to protect our intellectual property & limit the ability of generative AI developers to freely plunder our minds without permission or compensation.”
PROPOSED: We stop using the expression “AI.” It’s virtually meaningless at this point. Just marketing hype. Henceforth, everyone should just say what function is being automated. GenAI art would, e.g. be “the use of automation to generate new combinations of existing images.”
The NY Times argues “there is nothing ‘transformative’ about using [our] content without payment to create products that substitute for…it. Because the outputs..compete with & closely mimic the inputs used to train them, copying…is not fair use.”
Artists are poisoning AI image generators with Nightshade
“We created Nightshade because right now, AI companies hold all the cards—and we need to tip the power balance back in favor of artists.” 🔥🙏🏽⚖️
—
@ravenben
“Tech companies aren’t just stealing your your creative content—they’re making vast profits out of selling it back to you. AI-powered apps like ChatGPT are trained on vast amounts of stolen data. What’s worse, it’s all justified as a public good.”
“What he’s good at is taking other people’s work & putting his name on it, or doing stuff that you can’t check if it’s true.”
Wow, Emad personally possesses the very foundational attributes of generative AI. He steals & makes stuff up. Is he real?
And just like that Amazon threw hundreds of AI-narrated titles on the market.
Even though they are classified as machine narrated, this move will screw over authors who spend a lot of money on human narrators, and, of course, narrators most of all.
Why?
Well, those titles are…
THIS 👇🏽👇🏽👇🏽
Google says AI is actually “computational statistics,” and Meta/Facebook calls AI “a giant autocomplete system.”
Today isn’t even Wacky Wednesday, and I get to say…
I agree with Google & Facebook. 🎭
#IGF2023
#AI
:
@Kent_Walker
says "computational statistics" is a more accurate way to describe what we currently call AI.
@nickclegg
says AI is more like a "giant autocomplete system" as they're just guessing what the next word or token should be in response to a human prompt.
“OpenAI states that it & other defendants in a growing # of similar lawsuits “will ultimately prevail because no one—not even
@nytimes
—gets to monopolize facts or the rules of language.”
If OpenAI’s defense rests on this framing, they’re toast. 🤯
This. Really this. It’s oh so tiresome and ubiquitous. Please make it stop. Machines don’t learn. They are not inspired. They lack intelligence. They have no intentionality, no understanding. This is not to say they have no utility, but let’s have a honest conversation.
Really hope we put the whole “oh so a robot can’t look at art now but humans can?!?” debate to bed in 2024 because it’s based on zero self-awareness of how memory works.
“Anthropomorphizing image generators & describing them as merely being ‘inspired’ by their training data, like artists are inspired by other artists, is not only misguided but also harmful. Ascribing agency to image generators diminishes the complexity of human creativity.” 🔥⚖️
On that note, as per my previous tweet, here's a paper put together by industry experts and authorities on the topic of generative AI/Machine Learning and the impact this had on artists.
I suggest you read it and share.
“‘Training data’ isn’t the right phrase for what's used to train generative AI models—it dehumanises what is, in reality, human creative output. This isn't data - it's books, artwork, music & more.”
🙏🏽💙🙏🏽💙🙏🏽
It all started by calling creative works “content.”
#Discontent
'Training data' isn't the right phrase for what's used to train generative AI models - it dehumanises what is, in reality, human creative output. This isn't data - it's books, artwork, music, and more.
Calling it data makes it sound like it's just a low-value, anonymous input to…
How to say you have no idea what copyright is about without expressly saying it…
Note to all readers: copyright was designed—at least in part, to encourage creatives to share their works with the public without divesting their property interest. A central part of the bargain.
“Current LLMs are trained on text data that would take 20,000 years for a human to read.”
But we are told by apologists for theft that “AI learns the same way as humans.” Funny enough, I don’t know any human that’s lived to be 20,000 years old. Anyone?
Current LLMs are trained on text data that would take 20,000 years for a human to read.
And still, they haven't learned that if A is the same as B, then B is the same as A.
Humans get a lot smarter than that with comparatively little training data.
Even corvids, parrots, dogs,…
Dan calls on “everyone else who cares about the future of AI” to come to the defense of OpenAI & Microsoft. But he’s mistaking the present of AI for its future. We can do better than that. Don’t buy into a corrupt & blindered vision of our future. Let’s build the one we want.
The NY Times is asking that *ALL* LLMs trained on Times data be destroyed.
That includes GPT 3 and 4, Claude, Mistral, Llama/Llama 2 and pretty much any other model in existence.
I call on
@facebook
,
@cohere
,
@AnthropicAI
,
@MistralAI
,
@huggingface
,
@google
and everyone else…
“Artists are being exploited daily by having their work products and creations served up as ‘data fuel’ which has been mined and scraped from the web, without their permission, and through predatory data gathering practices.”
via
@Change
“These authors’ livelihoods derive from the works they create. But the Defendant’s LLMs endanger fiction writers’ ability to make a living…allowing anyone to generate—automatically & freely (or very cheaply)—text they’d otherwise pay writers to create.”
“Co’s & users are [using] GenAI to crank out high volumes of content. While the initial concern is abundance of content containing inaccuracies, gibberish & misinformation, the long-term effect is complete degradation of web content into useless garbage.”
“GenAI’s use isn’t’ transformative’…it is not commenting on or critiquing the original story or shifting it to a different medium, purpose or context. It is, instead, reproducing substantial parts of others’ work simply to compete in the same market channels as the original.”
@jmjafrx
@anildash
They don’t understand the nature of freedom in an interconnected world. They still cling to a two year old’s version: “You’re not the boss of me. I can do whatever I want.” They have failed to grasp that a sheep & a wolf won’t have a common definition of the meaning of freedom.
OpenAI’s response to the Copyright Office:
“By democratizing the capacity to create, AI tools will expand the quantity, diversity, & quality of creative works, in both the commercial & noncommercial spheres.”
Even OpenAI can’t really believe this. 🤯
BREAKING NEWS: The U.S. Copyright Office has been granted granted summary judgment against Dr. Stephen Thaler’s attempts to register GenerativeAI work!
A brilliant piece from Ted Chiang in
@NewYorker
:
“Obviously, no one can speak for all writers, but let me make the argument that starting with a blurry copy of unoriginal work isn’t a good way to create original work.”
Deepfakes aren’t just about celebrities. They affect each of us & erode an already fragile faith in truth in an ecosystem already battered by disinformation. What happened to Swift happens to women & girls every day. It is time to end this. Today.
An interesting & positive development in Japan. As the Japanese government is apparently moving forward with an ill-considered plan to allow unauthorized use of creative works to train AI,
@pixiv
, a leading online artist community site, chooses a different—more enlightened—path.
hey artists & writers, look what the cat dragged in. certain library associations have weighed in on AI, aligning themselves with our would-be tech overlords. 😡
@AuthorsGuild
@kortizart
“Any company adopting tools built on exploitative datasets are creating legal liabilities that will not fade into the background. "Everyone else is using it" will not absolve companies of responsibility.”
🙏🏽
@jovialjoy
In
#UnmaskingAI
, I write about the need for deep data deletion because there are core datasets (like LAION-5B) that are not only created without permission or compensation, but they are also composed of links to illegal imagery. Any company adopting tools built on exploitative…
@talkingdraft
@TheOnion
“But it was all so unpredictable. It never occurred to me that building a product by using raw materials I didn’t own or license was somehow wrong. I mean, seriously…who could have known that?”
@ProfNoahGian
I am so sorry Noah. The pain is unbearable. Here’s a photo of my Monty, a Great Pyrenees that raised my children better than I did. He lived to be 14, and died 12 years ago. There is still a hole in my heart…but my heart has also grown—perhaps three sizes. Much love to you. 💙
Also see: “Getty Images argued that ‘ask for forgiveness later’ opt-out mechanisms were ‘contrary to fundamental principles of copyright law, which require permission to be secured in advance.’”
@GettyImages
is 100% right. Opt-out is a formality prohibited by Berne & TRIPS.
⭐️⭐️⭐️⭐️⭐️⭐️⭐️⭐️⭐️⭐️
Wise counsel from the UK House of Lords—“we do not believe it is fair for tech firms to use rightholder data for commercial purposes without permission or compensation.”
h/t
@ednewtonrex
Sam Altman keeps on making pronouncements he doesn’t realize are indictments—
*the need to expand energy sources for AI
*massive investments of money for infrastructure
*the impossibility of operating AI if it requires consent for data harvesting, etc
For what—to lose our souls?
A “popular hype right now is AI art generators. Artists widely object to AI art because VC-funded companies are stealing their art & chopping it up for sale without paying the original creators.
Not paying creators is the only reason the VCs are funding AI art.”
—
@davidgerard
“The FTC has required businesses that unlawfully obtain consumer data to delete any products—including models & algorithms developed in whole or in part using that unlawfully obtained data. [We] will continue to ensure that firms are not reaping benefits from violating the law.”