The Quiet Theft: How AI is Consuming Creative Content
In the digital age, a new kind of silent crime is unfolding—one where the very essence of creativity is being quietly siphoned away by AI. Much like Napster disrupted the music industry, today's AI models are consuming vast amounts of copyrighted content, from songs and stories to artwork and ideas, often without credit or compensation. This "Napster of Everything" threatens not just the livelihoods of artists and creators but the integrity of creative work itself. It's time we take a stand to protect what makes our culture rich and diverse before it's too late.
In our family, I’ve done my best to teach our kids some hard truths. The first of which is that the best country music song ever written is “He Stopped Loving Her Today” by George Jones. The second truth is that the best modern country music song is “Whiskey Lullaby” by Alison Kraus and Brad Paisley. These facts are not open to debate.
Another fact that I don’t think is open to debate: AI companies are making billions of dollars using everyone’s copyrighted content. As an online publisher, we see this more than most. But it is not just a media problem. Let me tell you a story about a song…
Whiskey Lullaby was performed by Alison Kraus and Brad Paisley but written by the great Whispering Bill Anderson. How it came to be as a song is another story, but there is no doubt it has one of the greatest and saddest opening lines ever: “She put him out like the burning end of a midnight cigarette.” That is Nashville poetry.
So, as a bit more setup for why I’m writing this today, I joined my wife’s digital lifestyle magazine, StyleBlueprint, in 2020, just as the pandemic basically shut down our publication. My background is in technology and data, so I had a lot to learn about media and how to navigate business when most of your clients freeze all advertising.
We made it through some very hard times and have grown to 17 full-time team members and 20+ freelance writers. We know we aren’t huge in the grand scheme of media. If the New York Times in the publishing world is the entire Beatles music catalog, we are more like a hidden gem track from Florence + The Machine—we have a loyal fan base that deeply connects with what we do, but we’re not the global juggernaut.
Independent publishers like StyleBlueprint have a tough road. We are proud to be still standing. We have amazing readers and amazing advertising partners.
And, for full disclosure, we’ve done everything we can to embrace automation and generative AI that we can – except to have it write content. We believe people will always tell better, more relatable stories. However, AI is extraordinary at correcting basic grammar, executing drudgework automation, and helping solve technical problems. It’s a groundbreaking, generational advance forward in technology. The “rules” currently promote everyone using AI, and we utilize it. We just think the “rules” need to include compensation for everyone who contributes content that powers that amazing output others rely on.
We really don’t want AI to go away because it is mind-bogglingly good at what it does. But AI will collapse if we all don’t work together to deal with the elephant in the room—current AI “Large Language Models” (LLMs) are sucking up copyrighted content at an exponentially increasing rate. Like all content. Images, video, audio words, written words. Everything. In a nutshell, it is the “Napster of Everything.” For those of you who don’t remember (or, gasp, are too young to know), Napster was a peer-to-peer music-sharing program that almost killed the recording industry in the early days of the internet. It ironically exists in a new form post bankruptcy and is based here in Nashville, TN.
Think this doesn’t apply to you?
If you are a PR company, every release, case study, and client-facing marketing program has been (or is soon to be) analyzed, and your best ideas will be available – almost instantly – to everyone.
If you are a small business like a restaurant, every page on your website, including your menu, is now available to the world so that others can curate menus based on what others are doing at a pace that would have seemed unfathomable even two years ago.
If you are a venture capital or private equity firm, every client portfolio case study is being ingested, and your best ideas and approaches are being incorporated for everyone’s research and use.
If you are a wealth manager, investment banker, or anyone in professional services creating content for the benefit of your client and SEO, your best work is now being parsed, piecemealed, and spread across thousands of websites with no credit while at the same time diluting the results of your efforts.
If you have created content for any media, including print and television, your work is likely being added to training data – because your print or video usually ends up on the web. As an example, every magazine I know has some digital version (StyleBlueprint is digital-only, so we are more keenly aware of what is happening).
If you are a construction company or architecture firm, all of your public-facing work – and some of your private work – is being consumed, and you are 100% not being credited with any new or novel ideas you bring to the table.
If you are a software coder, you already know. Your excellent work is now automated for everyone else.
If you are a photographer, author, publisher, musician, songwriter, artist, actor, or screenwriter, you already double-know.
AI. Consumes. It. All.
The Napster of Everything.
As a prime example, in the last three years, OpenAI, the maker of ChatGPT that kicked off this revolution, has gone from $0 to $3.4B in revenue. Its most recent valuation round was $1T. Yes, with a “T” trillion. They will not disclose all the sources that go into their LLM. Some are public domain (think Wikipedia, which is fair use), some are licensed (they paid some money to Reddit, for example), and the rest are scraped or purchased from other sources.
Here’s an example, I asked ChatGPT to give me the lyrics of Whiskey Lullaby:
Pin
It started to show the song lyrics and then stopped because it knew it was a clear copyright violation. But the lyrics are still inside of ChatGPT and can be used to create additional work. Doesn’t seem like “Fair Use” to me. More on this in just a sec. Whispering Bill, you wrote a masterpiece. Just like a radio/streaming play, you deserve a penny or two.
So, how do we get out of this? We take a page from the Napster days:
There will be more lawsuits like the New York Times suing OpenAI and Microsoft. Their current case is compelling, but this will take years. If you recall, the Recording Industry Association of America (RIAA) sued Napster in 1999. It wasn’t decided until 2001. Recently, Eric Schmit, former CEO of Google, said the quiet part out loud about AI – Silicon Valley will take copyrighted content and let lawyers “clean up the mess.”
Until then, we are all on our own.
However, we can control what we allow the AI models to have for free from our websites. It’s a simple technical change any website owner can make. Here is a link to an article where I lay out instructions that you or your website manager can do right now at no cost to block access from generative AI web scraping. And you won’t be alone—over 36% of the top 1,000 websites already do this.
Just to make sure you got that, here is a link to an article where I lay out instructions that you or your website manager can implement TODAY to block access from generative AI web scraping. No one likes being told they should do something, but you should do this.PinChart from Originality.ai
Support some interesting startups that are trying to solve the problem proactively and hopefully without lawyers. Here are my two favorites so far:
ProRata.AI—Founded by Bill Gross (the guy who created pay-per-click advertising), ProRata is creating a publisher-friendly answer engine similar to Perplexity. However, ProRata will share all resulting revenue 50/50, and our content doesn’t go into general LLM training data. We are excited to join this effort along with many other publishers. Read more here about where they are headed.
TollBit—Another well-funded startup, Tollbit is creating a marketplace for any website to license its content to AI companies for training or real-time use. We love their approach and technology and are already signed up. They’ve been a joy to work with and have also given us some fascinating statistics about the overwhelming volume of AI bots attempting to hit our site (we filter bots from our traffic statistics).
PinChart of StyleBlueprint bot vs. human traffic from Tollbit
When the RIAA won its lawsuit against Napster, a whole ecosystem sprang up to help artists understand digital music streaming plays. Similarly, new startups like ProRata.AI and Tollbit are getting a jump on creating an accounting system for content creators (a broad term as it applies to a vast number of people in this use). To be clear, we are not expecting a big cash windfall by licensing our content. But every great journey starts with a first step. As an independent publisher, we just believe we have to try.
In summary, let’s keep the AI train rolling in a fair way for everyone. The genie is not going back into the bottle. We want to work positively toward content control and licensing solutions for everyone.
Now, with apologies to Whispering Bill, I had ChatGPT write a song about AI and content using Whiskey Lullaby as a base. It’s awful, but it proves that the original song lyrics are being used to create new ones in a commercial setting. There will be no big royalty checks for the following song. And if anyone were to get any cash, who would even get a cut? This is a problem that needs solving quickly before things go to far.
Song Title: “AI Lullaby”
Lyrics:
[Verse 1]
He was a writer, full of pride,
Crafted stories, deep inside,
But late one night, a silent crime,
AI took his words in the night.
[Chorus]
Singin’ “Hush now, don’t you cry,”
As the code begins to fly,
All his dreams and words, they fade away,
In an AI lullaby,
An AI lullaby.
[Verse 2]
She was an artist, bold and free,
Painted pictures, poetry,
But through the screen, it came to be,
AI stole her art, left her empty.
[Chorus]
Singin’ “Hush now, don’t you cry,”
As the code begins to fly,
All her colors, all her light,
Lost in an AI lullaby,
An AI lullaby.
[Bridge]
No one saw it come, no one knew,
A quiet theft in shades of blue,
But in the night, they whispered still,
Of the art that AI killed.
[Chorus]
Singin’ “Hush now, don’t you cry,”
As the code begins to fly,
All our voices, all our might,
Gone in an AI lullaby,
An AI lullaby.
[Outro]
Now the screens are silent, still,
No more stories, none to fill,
The emptiness we feel inside,
Just an AI lullaby,
An AI lullaby.
Our unofficial motto at StyleBlueprint is "Be kind. Do good." We encourage this to be the basis for all comments on our articles. Provide feedback that adds to the story. Some controversy or disagreements are part of any good dialogue between friends, but anything that tears down or belittles others is subject to disapproval or removal. Thank you for being a member of the StyleBlueprint community! View our Community Guidelines.
It started to show the song lyrics and then stopped because it knew it was a clear copyright violation. But the lyrics are still inside of ChatGPT and can be used to create additional work. Doesn’t seem like “Fair Use” to me. More on this in just a sec. Whispering Bill, you wrote a masterpiece. Just like a radio/streaming play, you deserve a penny or two.
So, how do we get out of this? We take a page from the Napster days:
There will be more lawsuits like the New York Times suing OpenAI and Microsoft. Their current case is compelling, but this will take years. If you recall, the Recording Industry Association of America (RIAA) sued Napster in 1999. It wasn’t decided until 2001. Recently, Eric Schmit, former CEO of Google, said the quiet part out loud about AI – Silicon Valley will take copyrighted content and let lawyers “clean up the mess.”
Until then, we are all on our own.
However, we can control what we allow the AI models to have for free from our websites. It’s a simple technical change any website owner can make. Here is a link to an article where I lay out instructions that you or your website manager can do right now at no cost to block access from generative AI web scraping. And you won’t be alone—over 36% of the top 1,000 websites already do this.
Just to make sure you got that, here is a link to an article where I lay out instructions that you or your website manager can implement TODAY to block access from generative AI web scraping. No one likes being told they should do something, but you should do this.
