Say NO to AI.
Authors woke up the other day to a new surprise. Meta’s platform, LibGen, stole god-knows-how-many proprietary books to train its AI. The dataset is 7.5 million but I don’t know how many books were in the public domain. Nevertheless, it is an enormous dataset of mostly purloined, pirated books. The new revelation, published in The Atlantic, comes after a similar article about Books3, another AI training dataset. The Atlantic’s article includes a searchable function that you can use to see if your works are among those stolen.
The Author’s Guild published a great summary article, which you can read here.
Personally, only one of my novellas is in the dataset, but the fact that even one exists is deeply disturbing. My Facebook feed overflowed with fellow authors decrying how many of their novels came up in search.
What’s All the Fuss?
I know that many people do not understand all the fuss. They shrug and say, “you have to move with the times. AI is the future.” But, let me ask you to think of these things:
- Our novels and novellas took us years to write, revise, and publish. These aren’t throwaway, discardable scraps. They’re our blood, sweat, and tears. They took time time and money, sometimes a lot of it, and we sell our work. It is a way we support ourselves and our families. Let’s say you were a professional chef and someone came in and stole all your handwritten recipes plus your cookware. Or, perhaps you’re a finance officer and someone marched into your office and stole your computer, hacked it open, and downloaded your work and proprietary software. Theft, right? It’s the same thing.
- Meta makes money off of our backs. Meta and OpenAI make a lot of money off of our work and not one penny flows to us, the authors. They pirated our work from existing pirate sites (pirating squared?) knowing full well that no one gave them permission to do so, used it to train their AI, and then sell access to that AI. They are making a fortune and we are told to shut up and sit down.
- They’re doing this to artists too. Let’s not forget the stolen images from actual artists. You’re an artist and you post pictures of your art on your website and whoops! AI bots scoop those right up and dump them into AI training protocols.
- The promise of AI is that Art, with a capital A, can be made by AI equal to or better than human Art, and completed faster. This is impossible. Art comes from the human mind. The human soul. Music, painting, sculpting, writing are all forms of storytelling and storytelling is one of the most basic human instincts we have. Long before the written word we told oral stories to warn our people of dangers and to pass on cultural beliefs. We painted on the walls of caves and made rudimentary instruments. We sang and danced and made Art, even from our earliest incarnations as humans. Computers, no matter how sophisticated, cannot do that. Whether they dream of electric sheep or not, it requires a human mind with an immortal soul to create something new.
Am I Against AI in All Forms?
No. Not at all. I believe AI can help us in many areas including science and medicine. It’s ability to process massive amounts of data makes it superior to the human mind for many such things. AI doesn’t belong in art, but can it help us elsewhere? I’m positive it can and will.
By the way, if someone had approached me asking to use my writing to train AI and had paid me for it, I might have said yes. But we weren’t given that opportunity because they snuck into our worlds and our words from the deep recesses of the internet and stole them out from under our pillows.