Have you ever wondered if you own the Kindle eBook you bought? Well to cut a long story short, you don’t. With a normal eBook downloaded onto kindle or a book with DRM enabled from another eBook provider you have a licence to read it tied to a specific account and that is all. You cannot lend an e-book to a friend after you have finished it, and god knows what would happen to your books if you lost your Amazon account for any reason.
However like most things digital there are ways around this, which can remove that restriction and give you full access to your eBook library locally. Also, with the typical e-reader having enough memory to store all the books you will ever read in a lifetime, you can then load them all on, turn off the WIFI and read any book, anytime without restriction. This also means you can back the collection up, so should your e-reader ever break or you get a newer model you can chuck the books straight back on and be up reading again straight away.
The problem is the usual dilemma for publishers and vendors, that there is a lot of trust required, trust that users will not break the copy protection and just email the book to a hundred people, or even worse, take the contents of the book and plagiarise it to create and sell an almost identical version. Both are possible, both are illegal and both are now done frequently, but good luck enforcing it.
With AI this latter may get worse before it gets better as the AI technically should not have access to copywrited material and therefore will not know it is breaking the law. When my last book (A Level of Consciousness) was just completed and as I was in the multiple read/edit stage, and I copied the whole of the book into Claude and asked the AI to show me any inconsistencies. It was an interesting exercise, giving me a list of totally illusional inconsistencies, and proving to me that AI models are still not up to the challenge of proof-reading, nor understanding the way humans think. However, it is only a matter of time until they are.
I am of the belief that the current AI’s, based on general purpose transformers iterating over large language models is an asymptotic curve, i.e. it will reach the sum total of human intelligence but not surpass it, but I also believe that is good enough. Train it better or give it more time (than the few seconds of free time it is giving you) and it will be able to do this and it could be a brilliant tool for writers who generally only find these things when somebody points out the obvious.
The big problem with this, is the terms and conditions of AI usage, that sometimes say that anything you give them becomes possible training data! So somebody asking the AI to write them a story about a dragon, may end up with something very similar to your book!
