Why ChatGPT Might Not Be Your Best Option for Book Recommendations

While ChatGPT's capabilities are impressive, there are more than a few reasons why it might not always be the best source for discovering your next favorite book.

In this blog post co-authored by the chatbot itself, I will even go as far as saying this is similar to a famous quote from the movie Jurassic Park—many people have become "so preoccupied with whether or not they could, that they didn't stop to think if they should" 😱

Limited Training Data

ChatGPT's knowledge is vast, but it's confined to the information it was trained on, which only goes up until a certain point in time. This limitation means that the latest books, emerging authors, and current literary trends might not be in its database. If you're someone who enjoys staying on top of the newest releases or exploring fresh literary voices, ChatGPT might not be able to satisfy your curiosity.

What's funny about this limitation is that it's the complete opposite of what talking to an IRL librarian would be like. Very often we can be your best source for what titles are flying off the shelves, have tons and tons of holds placed, or are suddenly very popular because of some other recent event.

Generic Recommendations

The charm of discovering a book that feels like it was written just for you can be lost when those recommendations are based on broad algorithms rather than nuanced understanding.

While ChatGPT can generate recommendations based on popular patterns and widely acknowledged classics, these suggestions can often come off as generic. Large language models have an issue with what's commonly referred to as regression towards the mean.

It's sad but true: your seemingly personalized results will always have a bias toward an "average answer" and in the process that might mean missing out on reading a true gem that a librarian knows you would like.

Missing the Human Touch

There's something special about book recommendations that come from a friend, a family member, or a trusted librarian. It's a feeling that is difficult to explain but happens all the time when you work at a library.

These recommendations are often accompanied by stories, personal anecdotes, or passionate explanations that an AI simply cannot replicate. "I just read this new book and think you will like it" adds value to the recommendation, making the book in question more than just a title, it's a shared experience.

No Accountability

When a friend recommends a book, there's a level of accountability and personal investment in your enjoyment of that book. With a chatbot like ChatGPT, there's no follow-up or shared experience to revisit. 

To put it another way, if the recommendation falls flat, there's no one to discuss your thoughts with, no one to understand why it wasn't a fit, and no one to refine future suggestions based on your feedback.

There are better options!

Now that I've used the classic sales tactic of telling you about a problem that you didn't know existed, here's how the library makes it all better 😉

Chatbots in their current form can be a great starting point for finding your next read, especially for those seeking quick suggestions or exploring a new genre. Without a doubt, they can help you get there, but for a truly personalized and engaging reading experience why not try the most trusted reading community out there: your public library. Here are some resources we offer at Palo Alto City Library that I suggest you try out before turning to a chatbot:

  • Shelf Made: Our in-house reading suggestions service.
  • Novelist Plus: An on-demand and expert curated book recommendations database.
  • BiblioCore Lists: Search similar titles and topic guides made by well read library lovers just like you!