Tech Bros Want Creativity for Them, Not You

Note: My regular Genre Grapevine column will be out at the end of the month. My previous coverage of generative AI has been collected in the free e-book Creativity in the Age of Machine Learning along with the follow-up report AI and the Enshittification of Life. If you like my genre reporting, please consider doing a paid Substack subscription.

When Meta pirated over 7 million books to train their Llama generative artificial intelligence model, the tech giant behind Facebook and Instagram also decided they shouldn't pay the books' authors because the company believed those books were worthless.

That is Meta's actual defense against a lawsuit brought by some of the authors of those pirated works, as reported by Keziah Weir in Vanity Fair. As Weir's article shows in detail, both Meta and the company's employees believed "individual books themselves are, essentially, worthless" because "none of Plaintiffs works has economic value, individually, as training data" for Meta's AI model.

Worthless. No economic value. Meta claimed this despite that tech company and many others being absolutely desperate for works created by humans on which to train their generative AI models. The irony is that "worthless" is what generative AI would be to tech companies without the ability to continually train these programs on books and other works created by writers, images and photographs created by artists, and songs and performances by musicians.

Big tech companies like Meta, OpenAI, and Google know this dirty little secret. And with artists, writers, musicians and many others increasingly fighting the theft of their copyrighted works by companies valued in the hundreds of billions of dollars, tech bros now fear losing access to their pirated gravy train. That's why people such as Jack Dorsey and Elon Musk are now saying we should "delete all IP law." Equally galling are the claims coming from OpenAI and Google, who want the U.S. government to let them train their AI models without restriction on copyrighted materials because it "is a matter of national security."

Welcome to our brave new world where your creativity might legally and permanently become a resource for others to use without payment and without your consent. Where the main worth of anything you create or do in your life might be in sustaining the AI systems created by incredibly rich and powerful corporations.

Welcome to a future where creativity is for them – for the tech bros, for the multinational corporations, for the elite, for the powerful – and not for you.

The Limits of Fair Use Claims

As Alex Reisner reported in The Atlantic, when Meta began development of their Llama 3 AI model, "they faced a simple ethical question. The program would need to be trained on a huge amount of high-quality writing to be competitive with products such as ChatGPT, and acquiring all of that text legally could take time. Should they just pirate it instead?"

Employees at Meta "spoke with multiple companies about licensing books and research papers, but they weren't thrilled with their options." These employees felt licensing the books and papers would be "unreasonably expensive" – never mind that according to Meta's own data, in 2024 the company had revenue of $164 billion and a net income or profit of $62 billion.

Meta employees also fretted over the "incredibly slow" process to legally receive those books and papers, which would take around four weeks. As a result, they decided to download and use a pirated database of books. The final decision to pirate was evidently made by none other than Meta CEO Mark Zuckerberg.

My novel Plague Birds was one of those 7 millions books pirated by Meta to train their Llama AI, as were other stories I've written. Because Meta pirated these books instead of licensing them, authors such as myself were not given any choice about our works being used in this manner.

But let's also be honest: Meta didn't decide to use pirated books because they couldn't afford to license them legally, or couldn't wait a few weeks to receive the data. The truth is the company simply didn't want to comply with copyright laws. As detailed in Reisner's Atlantic report, a Meta director of engineering summed up why the company went all-in on using pirated works: "If we license one single book, we won't be able to lean into fair use strategy."

Fair use is the legal strategy Meta and other tech companies such as OpenAI and Google are using as a defense against not receiving permission or paying for the right to use copyrighted works to train their AI models. Under fair use, you don't need permission to use a copyrighted work if you transform it into something that builds on the original in a different manner or exists for a different purpose than the original work. For example, parody of a copyrighted work counts as fair use, as does journalistic news coverage.

In the United States, federal courts are the ultimate arbiter of if something is fair use. To do this, the courts use a "four-factor test" that involves "looking at the reason behind the work, the nature of the work (whether it's poetry, nonfiction, private letters, et cetera), the amount of copyrighted work used, and how the use impacts the market value of the original."

Tech bros and tech companies are scared because multiple lawsuits contesting their claims of fair use have been filed by artists, authors, creators, and publishers. And earlier this year, a tech company lost the first major copyright case in the United States related to AI.

In this case, ROSS Intelligence was sued by Thomson Reuters, the company that publishes the legal research and law database Westlaw. ROSS evidently used aspects of Westlaw to train their company's AI model, which was "designed to provide a list of judicial opinions responsive to a legal-question prompt." As stated in this legal analysis of the decision, the judge ultimately ruled that ROSS's claim of fair use failed, in part, because ROSS was a for-profit company that would use their AI system for commercial use.

And in what was the biggest factor against ROSS's claim of fair use, the judge also said the fourth and final factor related to fair use, "which considers the effect on the value or potential market for the copyrighted work, favored Thomson Reuters" because "the court identified ROSS's AI tool as potential competition in the market for legal research platforms."

ROSS Intelligence has appealed the ruling against their company, but if it stands the case could indicate major trouble for other tech companies making similar fair use claims.

The fact that the judge in this case ruled against ROSS largely because the company was a potential competitor to Thomson Reuters is an interesting legal precedent. For example, the authors and writers suing OpenAI could make similar claims that OpenAI is a potential competitor. Last month OpenAI released a new AI model that is supposedly "good at creative writing," or so says the company's CEO Sam Altman. But whether or not the model is actually good, this new AI model definitely allows the authors and writers suing OpenAI to claim the company is their future potential competitor, possibly nuking any of the company's fair use claims.

The decision against ROSS Intelligence is also noteworthy because the judge "concluded that ROSS's use was not transformative, emphasizing the fact that the ROSS AI model is not generative AI, and that it instead 'spits back relevant judicial opinions that have already been written.'"

As I've written before, tech companies use a lot of deceptive language to describe what their AI systems are and can do. For example, these models are not truly "artificial intelligence" because there is no intelligence behind the systems, which instead use algorithms crafted from data samples, also called training data, to produce results. Despite this, tech companies continually use the term AI to promote their models because that grabs people's attention and brings in more funding, as opposed to the use of more accurate terms such as large language models (LLMs).

Because the ROSS case was filed in 2020, it's possible this company's AI is a different style of program from ChatGPT, Llama, and other current generative AIs models. Despite this, the legal argument that generative AI models return results "that have already been written" may still be a powerful one to use against the fair use claims of companies like Meta and OpenAI.

For instance, in March The Guardian published a short story about grief supposedly written by OpenAI's new creative writing model. Many people gushed over the story including Jeanette Winterson, whose essay about the story opened by saying "I think of AI as alternative intelligence."

Winterson was heavily criticized for this boosterism of AI copy, with many other authors and readers noting the obvious failures of the AI-crafted story. Nnedi Okorafor said the story was actually "stilted, heartless, and aggravatingly derivative." And it appears this derivative viewpoint is not merely Okorafor's opinion – as Gareth Southwell pointed out, one of the story's most eye-popping phrases, "democracy of ghosts," turned out to be copied from Vladimir Nabokov's 1957 novel Pnin.

Willow Roberts dug deeper into the AI story, which is about a "human trying to use AI to simulate conversations with a lost loved one." Roberts noted that since the story is "'metafictional,' it's really just the LLM talking about constructing such a story using borrowed human phrases from its data set."

All of this points to a potentially strong legal argument against companies using fair use to build their AIs, which is that generative AI systems are not truly transformative but instead merely borrowing or copying.

This argument may not be clear-cut in all the pending generative AI copyright cases. But as shared in a recent article in Forbes, tech entrepreneur and machine intelligence expert Chomba Bupe still believes this "undermines claims of fair use because AI models do not create truly novel content but rather recombine compressed versions of copyrighted materials."

What's Yours Must Be Ours to Use

The threat of fair use claims failing and being on the hook for massive legal payouts has tech companies and entitled tech bros looking for alternate ways to evade the intellectual property laws that protect artists, writers, and creators.

One way tech companies are trying to do this is to convince the U.S. government to allow them to train their AI models on copyrighted material. OpenAI, Google, and other tech companies recently submitted such proposals to the White House in response to Donald Trump's desire to craft an AI Action Plan that will "enhance America's position as an AI powerhouse."

I don't think individual artists, writers and creators will like what comes out of this AI plan, especially since Trump has stated he wanted the plan to keep "burdensome requirements" from impacting innovation. Combine that with how the tech industry heavily supported Trump both during and after the last election and it's likely he'll give the tech bros everything they want.

And what makes this change even more likely under the Trump administration is that these companies are connecting their arguments with national security concerns, saying if they don't get their way then China and Chinese companies will win the AI arms race. OpenAI is even using propaganda terms like "freedom-focused" and "freedom to learn" to describe what the AI industry wants from Trump. And other countries such as the United Kingdom have also been considering pro-AI changes to IP law, although artists, writers and musicians in the U.K. have pushed back hard on that.

But the tech bros don't stop there. On April 11, Jack Dorsey, the founder of the social media site formerly known as Twitter, posted the short phrase "delete all IP law." In response to Dorsey's post, X-Twitter's current owner Elon Musk replied and said, "I agree."

As Lincoln Michel ‪noted in response to this, "none of Jack or Elon's companies would exist without IP law. Neither of them are offering to give up their patents, trademarks, or other IP. They just hate artists."

I agree with Michel that Dorsey and Musk's companies and their own billions in wealth would not exist without IP law and copyright. Dorsey and Musk are also definitely major hypocrites in wanting to pull up the IP ladder after they climbed to the very top. But I also believe this is about more than just hating artists – Dorsey, Musk, and others like them appear to be trying to rebrand creativity itself. Just as tech companies and their backers have used loaded words like artificial intelligence to describe LLMs, they want to change how people see creativity itself so that it better benefits those with power, money and connections.

Some of what Dorsey said after his "delete all IP law" statement seems to support this rebranding effort. For example, when Nicole Shanahan disputed Dorsey's statement about getting rid of IP law, saying "IP law is the only thing separating human creations from AI creations," Dorsey responded by saying "creativity is what currently separates us, and the current system is limiting that, and putting the payments disbursement into the hands of gatekeepers who aren't paying out fairly."

It's bad enough that Dorsey appears to believe that artists, writers and creators – along with the publishers and market systems used to share our work – are nothing more than "gatekeepers." But worse, he is saying that creativity is all that separates humans from AI and that we need to find a way to change that.

In a later tweet, Dorsey added that "speed and execution matter more" than anything else with regards to creativity.

Looking at all of Dorsey's recent statements on this topic, I see this as an attempt to rebrand how people see creativity. Instead of creativity being about finding new and original ways to see the world, or building something that resonates with people, or doing something that connects people with each other, or any of countless other variations on what it means to be creative, the tech bros want creativity to be mainly seen by humanity as being about speed and execution.

The tech elite know about the limitations of their generative AI models. They know their so-called AIs don't actually create anything new or unique. Instead, the algorithms in their systems turn out copies and imitations of what has been done before. They pull together derivative variations on what people have already created, said, and done.

But the main things going for their AI models is they are far faster than any human will ever be, and the execution looks professional enough that a lot of people accept the results.

Just as the tech industry has used words such as artificial intelligence and freedom-focused to push their agenda, now I fear they want to change what creativity means to our lives to better fit the needs of their companies.

Creating a New Future

But what does all this mean for anyone who isn't a tech bro billionaire?

First of all, the tech companies running these AI models absolutely need continual access to the fruits of human creativity to keep advancing their systems. These systems only produce decent results when they are continually trained on works created by actual humans. In 2023, BBC's Science Focus published a must-read article about the hidden echo chamber powering AI models. The article quoted Ahmed Elgammal, a professor of computer science at Rutgers University, who said, "If you get into the cycle of feeding (AI models) what is on the internet, which right now is mostly AI, that will lead to a stagnation where it is looking at the same thing, the same art style, over and over again."

In simple terms, this is like a never-ending cycle of making a photocopy of a copy of copy. Very quickly this cycle produces extremely degraded results. Essentially, if generative AI isn't continually trained on works created by actual humans, the programs stagnate.

This concern is likely a major reason why OpenAI's recent policy recommendations to the Trump administration said "If the PRC's developers have unfettered access to data and American companies are left without fair use access, the race for AI is effectively over."

Societies around the world have long devalued creative work by artists, writers and others. Creative fields frequently pay relatively little unless you're at the very top or very lucky. But now that works created by humans are desperately needed by the world's richest people and the world's most powerful tech companies – now that writers, artists and others are pushing back on the theft of their works – the tech bros and their companies want to rewrite the rules so they can keep stealing anything they desire.

That's why tech bros call us gatekeepers. It's an attempt to shift the narrative from what they are doing to the victims of their actions. It's sleight-of-hand, to distract from them needing us to keep being creative so they can feed our creations to their ever-growing AI models.

Pay attention to what our tech overlords are doing and how they frame their actions. They want to create a world where anything of yours can be stolen. And yes, in this case "yours" means all of us, not merely the artists, writers, and musicians fighting back right now. The tech companies eventually want their AI models to take over vast parts of the economy. Their next targets will be the jobs of lawyers, doctors, farmers, nurses, accountants, managers, production workers, and many, many others. People currently working in creative industries are merely the canaries in the coal mine with regards to what tech companies plan for the coming years.

The ultimate goal of our so-called tech overlords is for the world to embrace a new technological feudalism. A world where unless you are extremely rich, nothing in your life is owned or controlled by you. Not your stories, not your art, not your home, not your work, not your time, and perhaps not even your life.

This is the argument that Meta essentially made defending themselves for pirating more than 7 million books. These tech companies don't see anything you do as an individual as having value.

That's why they want to change our laws. Human societies long ago decided through our laws that theft isn't mitigated by stealing so much that any individual item you stole is worth very little compared to the whole. Or as Lincoln Michel said, "Official tech corp position is that books are so high value that it would be impossible to train quality LLMs without them AND books have zero value and thus it is ridiculous to expect global tech corps to pay any money for them."

Tech bros know that their AI models are unlikely to create anything that breaks the mold. That their AIs will struggle to create anything that pushes the bounds of human experience, or challenges the powers that be. Instead, the creations of their AIs will be, by design, copies and imitations of what went before. It will be middling, but nothing that changes the world.

But their vision of creativity will be very, very fast. And it will make their companies untold trillions of dollars. And to reach and sustain that future, they need continual access to everything humans have ever created and will ever create.

I'm still irritated that Meta trained their AI on my novel Plague Birds. But I'm also amused because the novel is about advanced technologies such artificial intelligence causing massive upheavals in both our lives and our world.

I know Meta's generative AI didn't understand my novel, because there is no intelligence or understanding within the program's algorithms. But if it had, I wonder what the AI would have thought of the story. Because Plague Birds doesn't merely focus on AI and other technologies devastating our world – my novel also focuses on what comes after that. That humanity, no matter how changed from what we are today, will always keep trying to find meaning and happiness both in our own lives and in our connections between each other.

At its core, my novel shows that by working together, humans can achieve anything.

There is a long history of the rich and powerful dismissing the actions of individual people as worthless. But you know what's actually worthless? Not even trying to create a better future.

When we stand together, no matter if it's as artists, writers, musicians, or as all people everywhere, we can do anything. We don't have to accept the tech bros' vision for our future and our lives. We can change that future to one where what we do and create as individuals matters.

Instead of being seen as worthless, we need to show these tech fools that it's they who need us. That it's they who can't exist without what we create.

There's nothing worthless in such a goal.