For those of you who are new to this topic, this is NOT ABOUT STOPPING CHAT GPT or using it in your work. IMO, go on and enjoy Chat GPT! This is about billion-dollar AI companies compensating the authors who created the conversational language aspects of the system… and have not yet been compensated for their work (spoiler alert: if there are pirated copies of your books on the web, it’s probably you.) 

This is also about folks in Hollywood who are now on strike because they fear that what’s happened to the writing community may happen to them as well. 

I’ve attached an updated blog post I’m working on for reference and information. I hope it helps. 

UPDATED POST: We need to talk about ChatGPT and similar AI. 

Long story short, the ‘talking side’ of these systems are driven by a language set of books that were stolen from dark web pirate sites[1]. The fact that this is theft was essentially confirmed by a recent Supreme Court (SCOTUS) decision[2]. The related class action lawsuits have begun. 

Let me explain. 

I’m the author of 45+ books of epic fantasy and science fiction. But in a previous life, I co-wrote an AI patent with Dr Constantine Papageorgiou from MIT’s AI lab. I’m also studying for my Masters’ Degree in History from Harvard, which helps me further understand complex systems. 

Chat GPT is a predictive model for language. Imagine a predictive model for weather without any weather history to predict from: It would be worthless. And $0 is how much value Chat GPT would have without a large corpus of long-form stolen content—again, much of which is pirated books–as its engine for conversational language. SCOTUS agrees that you can’t take someone’s IP in order to directly compete with them, even if the results are only served up in ‘fair use’ chunks of less than forty words. 

AI companies like Chat GPT are perpetrating the greatest theft of intellectual property in history, period. Chat GPT (one of many AI companies) just got $10B from Microsoft for one partial investment (one of many to come). If you assume Chat GPT didn’t pay for the 300K books it stole from the dark web, that’s 30K/book/author that was denied to the people who trained the system… and again, ChatGPT is just one of many AI systems using the same stolen dataset. 

Personally, I played around with Chat GPT. At first, I thought it was funny how the system tried (and failed!) to copy my voice and characters. It was only later on that I wondered how it got enough info to even try in the first place. In other words, I used to work in AI and it took me a while to realize what these folks were doing. 

And the class action lawsuits have just begun. Here’s a description of first two lawsuits from Rolling Stone: https://www.rollingstone.com/culture/culture-features/chatgtp-openai-lawsuits-copyright-artificial-intelligence-1234780855/ 

I filled out the form in hopes of joining the Saveri class action suit: https://www.saverilawfirm.com/chatgpt-language-model-litigation

I also filled out the form in hopes of joining the Clarkson suit: https://clarksonlawfirm.com/togetheronai/ 

The author’s guild is putting forward a letter to ask for compensation. IMO, this will probably turn into another lawsuit, so I joined this as well: https://authorsguild.org/news/sign-our-open-letter-to-generative-ai-leaders/

Common Questions: 

Won’t lawsuits put ChatGPT out of business? 
By the time the AI wave ends, billions will be put into AI companies. They are not going away just because they have to pay a fair wage to the people who made them rich the first place. 

But I like ChatGPT. 
Thank you. It’s my words–and those of thousands of other authors—that you enjoy. 

How can ChatGPT really know who wrote what bits of a response? 
The system serves up the next most likely statement, which is like a chef serving up a meal. A kitchen can’t consistently serve good onion soup if one grocery store always delivers terrible onions. By definition, you need to know where good content is coming from and act accordingly. It’s all in the system and lawsuit discovery will turn up the truth. 

What does this have to do with the Hollywood strike?
Hollywood writers are concerned that studios will use a loophole in current contracts to have ChatGPT ‘write’ a first draft… so they only get paid for revisions, which is less $$ and also no residuals. On-screen talent is worried that studios will create deep fake versions of them to use before and after their death (without compensation.) The studios have been extreme in saying they will not add any contract language that prevents them from using AI in any way they wish. And since he makes $31M/year, Bob Iger earns more in a day than the average writer does in a year. Long story short, studios can afford to pay people for their likenesses and work, period.

What can I do to help? 
Feel free to keep using Chat GPT if you like it. But please help raise awareness that AI companies need to pay those they stole from. 

Who’s really joining these class action lawsuits? 
Most recently, Sarah Silverman jumped into the mix, which is helping to raise awareness. Thank you, Sarah! https://www.nytimes.com/2023/07/10/arts/sarah-silverman-lawsuit-openai-meta.html 

Also, the FTC recently opened an inquiry. https://www.nytimes.com/2023/07/13/technology/chatgpt-investigation-ftc-openai.html 

Should I file for a copyright on my books? 
I’m no lawyer, but I’m certainly filing. Instructions are here: https://blog.reedsy.com/guide/copyright-a-book/ 

This is BS. Where’s your proof? 
The sources for my statements above are in the footnotes below. 

[1] Jack Brandy, “Dirty Secrets of BookCorpus, a Key Dataset in Machine Learning,” TowardDataScience.com, May 12, 2021, https://towardsdatascience.com/dirty-secrets-of-bookcorpus-a-key-dataset-in-machine-learning-6ee2927e8650 

[2] “U.S. Supreme Court Agrees with Authors Guild in Fair Use Case,” AuthorsGuild.org, June 6, 2023, 
https://authorsguild.org/news/us-supreme-court-agrees-with-authors-guild-in-fair-use-case/