From social media posts, to songs, to even pieces of art, generative AI has been a game-changer. But as most know, generative AI’s inputs aren’t unique, and are, in fact, acquired from massive amounts of training date. It is this training date that provides AI with its ability to create various content upon request. Using Large Language Models (LLM), these systems can leverage tremendous volumes of data to generate works in seconds. But therein lies the rub. Increasingly, generative AI is being used to create published materials that are essentially poor reproductions of existing works. And the latest involves fake books on Amazon that are little more than awkwardly paraphrased than the originals.
(Photography is an art–will generative AI kill it? Read this Bold story and find out.)
These new AI book scams are everywhere and are quickly getting out of hand. In some cases, AI-created new books are being promoted under an actual author’s name without their knowledge. In other instances, AI is being used to list fake books on Amazon that copycat existing ones. While it’s evident that this is clearly wrong, it seems real authors and publishers lack an easy way to combat their occurrence. It remains unclear whether AI content represents copyright infringement or not. And AI-generated materials can be difficult to detect when dealing with large volumes of works. As such, it would appear that, for the moment, AI has indeed gone rogue. And ultimately, it may be the consumer who suffers the most.
The Latest Evolution of Cliff Notes
Since the dawn of publishing, authors have had to fend off would-be copycat writers. Imitation works have always been a bit of an issue. But in the past, these works were actually created by humans, not by a machine. Today, the fake books on Amazon and other online sites are just the opposite. Of course, an individual or even a shady publisher is behind these efforts. But the actual content of these books is generated by AI and its use of LLM. Not only does this raise new issues about copyright violations. But it also makes it more difficult to monitor and police. Since this new AI book scam can populate sites with thousands of works within days, things are notably challenging. And by the time a real author realizing their book has become an AI victim, hundreds of consumers may have already made a purchase.
(Generative AI is reshaping education–read up on it in this Bold story.)
In most cases, the quality of these fake books on Amazon created by AI is extremely substandard. But for some readers, that’s ok. This is especially true for students or others who are looking for a quick summary of the real thing. Summaries of existing works can be quite beneficial from a time efficiency perspective. That’s why Cliff Notes as a company has done well for decades. But when the summaries are almost direct copy-and-pasted passages rather than a well-analyzed summation, everyone loses. Real authors credibility can be undermined if readers believe the fake books on Amazon are legit. And readers suffer this new AI book scam as they waste the money. These scams may not be new in total, but the ability of AI to produce these in high volumes so quickly is a game-changer.
Identity Theft in the Publishing World
It’s certainly bad enough to have one’s creative works diluted with a barrage of fake books on Amazon. This new AI book scam results in potential customers not knowing which is real and which is fake. But imagine having AI-generated books listed on these online sites under your own name. That’s precisely what has happened to some well-known authors like Jane Friedman. She recently began seeing several titles of books listed under her name on Amazon. And when she went to examine them, it was clear they were AI-generated content mimicking her actual works. Amazon has no clear policy to verify authorship either, which could help address such a situation.
In this instance of fake books on Amazon, the issue isn’t really one of copyright infringement. Indeed, Jane Friedman’s materials are being used to train AI systems on how to write like her. This could be viewed as a violation of intellectual property. More importantly, the AI-generated books were using her name and likeness to recruit customers. This is more of a right-to-publicity issue itself, which reflects an entirely new AI book scam. Whether it’s dilution of one’s works with poor AI content or wrongfully leveraging one’s reputation, things are out-of-hand. And the options to address these developments are both limited and costly for those wishing to fight the system.
Controlling the AI Chaos
When it comes to dealing with fake books on Amazon, there are many stumbling blocks at the present time. From a legal perspective, it is not yet clear whether or not AI-generated content violates copyright laws. Because generative AI trains on others’ works, in this case authors’ books, there could be support for this. This is why the New York Times recently filed suit for copyright infringement against OpenAI. Similar to the current new AI book scam, AI was producing large segments of published works verbatim. As such, it may be that legal protections could exist. But even if they do, these aren’t likely to be available anytime soon. And even once in place, it would still require legal actions for authors and publishers to go after these scammers.
Alternatively, Amazon and other online publishers could choose to do a better job of screening for AI-produced content. This doesn’t mean banning such works, but it might involve requiring transparency and disclosure. Currently, it remains the reader’s responsibility to determine whether they’re buying fake books on Amazon. Often, they may not realize it until after the purchase. Likewise, authors tend to be the ones to bring it to Amazon’s attention rather that the other way around. Understandably, the task for Amazon would be herculean given the volumes of book uploads they receive. But having some policy and practice in place to reduce this new AI book scam would be helpful. At present, this remains lacking, which is why AI-generated books will continue to plague online sites for the foreseeable future.
