As it is now, the personalized recommendations seem to pull from your average ratings to determine what is a low or high score. However, I am more likely to choose a book I think I will like, so of course most of my books are rated 3 stars or above.

This becomes a problem when the AI says something like "This book has _____ theme, which is present in your lower rated reads (Title, 3.75 stars)"

If we could establish a custom (or even if a general one was established) scale for what our ratings mean the AI could compare it to that instead, which would be helpful.