You could make a working version and then try to make it cheaper.
Do you want to store numbers with the spoken voice on the Arduino ?
I suggest to use a Raspberry Pi Zero with a SD memory card to store the data. With a script, for example Python.
You need a USB adapter for a speaker and a microphone.
Inputs from where? Directly from the page with a camera?
It is supposed to read out loud, I get, but do you also want to record that too?
It's good to have a goal, but that may change. First make it work, then make it less expensive.
There are many videos on youtube of people making OCR (optical character recognition) projects and image recognition projects with their Arduinos. Start your research there. There are plenty of free resources just waiting for you to use them.
Absolutely is, using AI it can be done offline, it can also be done with an internet connection.
The Nano is a good place to start, but you should have some kind of plan before investing since money is a concern for your project.
As you can tell from the responses you got, this is not so much an engineering question as more of an entrepreneurial one that encompasses a wide range of topics.
I've done a lot of teaching, coaching etc. of tech-entrepreneurs. I always make a distinction between those that have an idea, and those that have just a dream. Yours is still just a dream. To bring it to the idea-stage, you'll need to figure out some of the key aspects of both the technology and the business side.
Ok, cool. It's starting to come into focus.
You want to read a text file containing a book, out loud from an Arduino?
A tiny amp and speaker, headphones, bt, whatever for output
SD reader and/or bluetooth
Ok, a suggestions from a different angle, then. How about teaming up with somebody who shares your passion (or, whom you can convince about yours :)) but with more extensive programming experience and/or knowledge of embedded systems? It sounds to me like you're the guy with the dream and potentially the drive to make it happen - you can use that 'get up and go' spirit to stimulate people around you to realize the necessary building blocks. This approach would prevent you from getting bogged down into the rather boring complexities of embedded systems engineering so you can focus on the actual goal at hand.