Artificial Intelligence (AI) has rapidly advanced in recent years, and its use continues to grow. For libraries, understanding its applications—and the potential risks—is essential, especially when it comes to collection development. AI is being used by authors and publishers alike to streamline their work. This can be problematic depending on the application and whether that information is disclosed or not. Content created by AI is not necessarily bad, but as library professionals, we should evaluate all materials carefully to ensure we know all relevant information before adding them to the collection. Patrons will sometimes request an item that is AI-generated without knowing, because they are interested in the subject. We’ve compiled some information to help libraries determine if books are AI (they can be quite well disguised) and decide if those items should be added to their collections.
Risks of AI-generated content
Copyright
AI platforms have been “trained” with real content made by real people. Whether it be images, articles, or books, AI is spitting out content that is not truly original and without giving credit that’s due. Read this article on a recent AI copyright infringement lawsuit to get a deeper understanding of the ramifications.
Anthropic pays authors $1.5 billion to settle copyright infringement lawsuit : NPR
Misinformation
I’m sure anyone who has tried an AI platform is aware that they do not always provide accurate information. For fiction works, this isn’t necessarily a problem, but there are plenty of AI-written non-fiction works that are not verified, reviewed by a publisher, or have much oversight at all. An AI-written medical book, for example, could be very harmful to patrons. As library professionals, it’s important to keep this content out of our collections to maintain the library’s trustworthiness and reputation of being a source for accurate information for Albertans.
How to tell if a book is created with AI:
- Self-published items have a higher chance of using AI. Publishing houses are responsible for reviewing the materials they put out and will often spot the use of AI before the book goes out for purchase.
- If the item has images, this is often the easiest way to tell. Check fingers and toes, cultural stereotypes, and other generally uncanny features of the illustrations.
- Has the author promoted the book release? There have been cases of AI-published books under well-known author names when the author has not signed off.
- If the book only has reviews in online stores and Goodreads, this may be a red flag.
- Copyright pages, index, and acknowledgements are frequently left out or simplified in AI materials.
Keeping AI in mind for your collection development policy
Your existing collection development policy will help you to evaluate AI materials just as you would an original work. Points to consider are quality control and transparency to patrons. Some libraries have decided to exclude AI-generated content altogether. Often, quality control measures that address self-published books would also be applicable to AI content. Many patrons will want to be aware if materials are AI-generated, and you may want to consider differentiating them in some way.
Further Reading:
Addressing AI-Generated Materials in the Library Collection – The Digital Librarian
5 Easy Ways To Tell If Written Content Came From Generative AI
Did a Robot Write This? Tips for Identifying eBooks Written by AI | Pima County Public Library