HOPPING mad

A writing colleague posted on social media that The Atlantic Monthly had done an in-depth story about Meta illegally uploading thousands of copyrighted works into their new Artificial Intelligence project, Llama.

The Atlantic Monthly very kindly left their searchable database outside the site's paywall, meaning that non-subscribers can check listings.

So I did.

And poof! just like that, I am now part of a class-action suit against Meta.

But why?  It's a little complicated, but first an overview of copyright.  ©
​
​Copyright, as we have defined it since the 1970's, is the innate right of any creator to control what happens with their creation (book, song, drawing, movie, sculpture, all the cool stuff) for the lifetime of the creator plus 70 years.  

The idea being that if you've gone to the effort to make something, you should be able to profit from and control (to some extent) what happens to it for long enough for you and your heirs to benefit.  

Which is why, if someone wants to make a movie of your novel, they need to obtain the movie rights from you.  You can make a deal or not make one, depending if you like the cut of their jib.

Or for example, the fellas in Aerosmith can say no (or send a cease-and-desist) to some rando using "Walk This Way" at a racist pep rally.  If those jerks want a rally-song today, they should turn to something older, like, say, 1923's "Yes We Have No Bananas" by Forman and Nash, or better yet, find a likeminded collaborator and compensate the artist for the use of the work.

©

​There are a few exceptions to copyright, but they are small and specific: anyone can quote a small part of of copyrighted work, for instance, for the purposes of review or scholarship. 

We all use copyrighted material, assuming we read books, buy artwork, listen to modern performances of music, watch K-drama, etc., etc. And we mostly pay fair value for the pleasure: buying a copy of a book, paying taxes that support libraries, purchasing an entry ticket, keeping up on subscription fees, or even renting our ears and eyes to advertisers by the 15-second-increment.

The wicket grows sticky almost instantly when the user of the copyrighted material puts that original work to use for gain.  Like when Vanilla Ice five-finger-discounted the "Under Pressure" rift from Queen and David Bowie for "Ice Ice Baby." ​

Vanilla Ice, bless his goofy heart, ended up having to apologize, give credit to David Bowie and Queen, and pay for the use of the riff, btw.  See also "cautionary tale."

And that's copyright generally sorted. Now the Artificial Intelligence (AI) kerfuffle.

Say Jane Doe, avid reader and would-be novelist, buys a copy of Novel X and reads it, later crediting the influence of that novel in her own success as a novelist. How is that different from Llama AI gobbling Novel X up in an effort to learn how to write and speak? ​

Such a good question. I have opinions!

To start with, what would AI be doing with Novel X anyhow? Overly simple explanation: the way that these AIs like ChatGPT, Llama, and even Grammarly learn language is through examples. Give it Shakespeare, and you get Shakespearean English. Which is not helpful when you want to make your hastily written e-mail sound more businesslike. Alack! A pox on your deadline, sir!

So in order to be really useful, AIs need an ocean of source material. Like, say, a library.  Only thing is: copyright.  ©

To get hold of Novel X, someone somewhere somehow in the AI world has to upload (which is to say made an electronic copy) the work. 

For instance, an impatient fan got the clever idea of having ChatGPT finish the last book in the Game of Thrones series in author George RR Martin's style. Except of course, George RR Martin holds the copyright to those books, and he never gave permission for the series to be electronically copied to ChatGPT.

​Cue the lawsuits.  ​​©

From "Mother Goose's Melodies" Copyright 1833.

So, back to Meta, Llama, and my very own brainchild novel, and why I am so dag-nabbed angry I have moved beyond actual swearwords.

And it's not because I am concerned that someone will have ChatGPT write a fan-fiction of my novel. No, it's the unseemly and unscrupulous uses of my work that bug the stuffing out of me.

First, according to the research done by The Atlantic, it looks like Meta lifted literal millions of books from a pirate site.  Let that sink in: instead of paying me the paltry $7.99 for an e-book, the scamps at Meta likely went to the black market.  For shame!

A company posting 1.64 BILLION dollars a year in revenue didn't even pay market value for source material? 

Second, they seem to have scrubbed the copyright notice (that page in the front that says where and when a work is published and also, not-so-ironically, who is the actual owner of the copyright) from these works.  Which speaks to willful decisions being made—and an actual understanding that what they were doing was in violation of copyright law. ©

Third, and I suspect this will be the sticking point legally, Llama AI is not a human, working on creating an original work of her own; it's a for-profit business product that stands to make bajillions of dollars for Meta as well as for users of Llama, a list that currently includes Google, Goldman Sachs, etc etc. More than 170 million downloads, according to Meta's own website.  

Bajillions of dollars, let me reiterate, off the backs of many, many writers like me.

And what do we get?
Zip, and absolutely no creative control when someone wants to imitate us. A bad actor could easily say anything in imitation of my narrator's voice.  Like, "Hey Google, write me a 30-second TV advertisement script in the tone of Margaret Atwood's The Testament for a new non-hormonal drug regimen for ED."

Which is horrifying and exactly the opposite of creative control. ​​

The situation feels very genie-out-of-bottle-ish: it's not like Llama AI can magically and easily excise the stolen property it may be made of (can it?), but I sincerely doubt anyone is going to throw the project out.  

So, like so many thing, the issue goes to the courts to decide how/if to settle writers' complaints. And meanwhile, the situation extends beyond the US. Writers in France just filed a lawsuit, and I suspect it's going to avalanche from there. ©

REFERENCES
https://routenote.com/blog/all-the-songs-now-in-the-public-domain-2024/

https://www.copyright.gov/what-is-copyright/

https://abounaja.com/blog/copyright-infringement-cases
https://www.txpatentattorney.com/blog/common-copyright-violations-you-probably-commit/

https://www.socialmediatoday.com/news/meta-court-trial-over-use-of-copyright-material-ai/741985/

https://authorsguild.org/news/meta-libgen-ai-training-book-heist-what-authors-need-to-know/

https://techcrunch.com/2025/03/21/meta-has-revenue-sharing-agreements-with-llama-ai-model-hosts-filing-reveals/

​https://www.statista.com/statistics/277229/facebooks-annual-revenue-and-net-income/

​https://about.fb.com/news/2024/05/how-companies-are-using-meta-llama/

https://originality.ai/blog/how-do-ai-writers-work#:~:text=Essentially%2C%20an%20AI%20content%20generator,content%20that%20matches%20the%20prompt.

https://www.reddit.com/r/technicalwriting/comments/1ddr4tz/tech_writers_how_do_you_use_ai/

https://wjlta.com/2024/03/05/plot-twist-understanding-the-authors-guild-v-openai-inc-complaint/

https://screenrant.com/game-thrones-books-ai-completed-removed-lawsuit/