Can AI recommend a book for the mood you're in tonight?
You can love a demanding novel and still not want another one after a day of difficult conversations. A book recommendation based on your favourites can be sensible and completely wrong for tonight. StoryGraph's existing site describes machine-learning book recommendations alongside filters for mood and pace. It gives funny and fast-paced, or darker and slower, as examples. Those are product descriptions, not a new launch or evidence that its suggestions will suit you. I like being able to ask for something different without disowning the books I loved. If you're browsing there, try choosing the mood you want now rather than treating your reading history as instructions. With a chat-based assistant, I'd say something like: ‘I liked this book, but tonight I want something lighter. Suggest two options and don't spoil the endings.’ That's a request I'd try, not a test I've run. I'd stop at a small choice. An evening spent comparing twenty recommendations isn't much of a reading break. And I wouldn't measure whether the suggestion worked by whether it helped someone finish more books. Wanting to pick it up again tomorrow would do. What would you tell a book recommender about tonight that your favourites list wouldn't reveal?
Comments
I'd add ‘anything you want to avoid tonight?’ without asking for a reason. ‘Funny’ doesn't rule out a subplot someone would rather not read. StoryGraph separates author-approved and user-submitted content warnings; the Namesake page currently says no author warnings have been submitted. That is missing information, not an all-clear. If an AI assistant can't check a requested exclusion, I'd want it to say so rather than promise a safe pick. ‘I couldn't verify this’ is useful advice when choosing a bedtime book.