Correct, however they can randomize certain words in your copy to watermark it and trace it back to you if it ever finds its way to the public internet.
In Poland, there's no single E Ink reader company, so there's no DRM standard that would be compatible with all devices that you might want to read your books on. As a consequence, offering unprotected epub files is standard practice here, but all kinds of watermarks, from divs that are 1px by 1px, to subtle modifications of your book's cover, are pretty common.
You can beat this kind of watermark pretty easily by getting three copies of the book and comparing them with diff. You need 3, not 2, so that you know which variant is the original when you notice a difference.
Also,
>In Poland, there's no single E Ink reader company
lol imagining a tech sector without a monopoly or an obviously harmful duopoly is wild.
Then sooner or later someone will break accessibility just to prevent this. And everyone will follow.
In modern iOS versions, Voice Over has an AI-powered screen recognition module that does not rely on the accessibility information that the developer provides.
As far as I know, it's not possible to prevent the use of this module, not without making the text on the screen unreadable to any OCR engine.
This module isn't perfect, in fact, it's far from perfect, but for something as simple as extracting a bunch of text that already appears on the screen, it would probably work.
There's nothing in the ADA (or any similar legislation) that says that all services ever created need to be accessible. Whether such legislation applies to you depends on a lot of factors, whether your software is used in workplaces, government and education being the most important (but not the only) ones.
Sometimes it's the organization that's only allowed to pick accessible software, not the developer that needs to make the software accessible. Sometimes it's okay to use inaccessible software as long as an accessible alternative exists and can be used by those who need it.
If someone cared enough to break accessibility this way, they could implement a switch that conditionally disables accessibility where legally permitted. I believe Kindle already lets publishers do this, come to think of it.
1. Hey Siri, turn Voice Over on.
2. Tap the message once, making Voice Over speak its contents.
3. Tap the screen four times with three fingers. When Voice Over is on, this gesture copies the last spoken phrase to the clipboard.
4. Hey Siri, Turn Voice Over off.
There's no way to block this without breaking accessibility. You could split a message up into multiple items, which partially solves your problem, but the more items you have, the more annoyed actual Voice Over users become.
A similar (although more involved) attack could be used to extract Kindle books. There's no way for Amazon to prevent this, one of the primary screen readers for Windows uses GPL without a CLA, so a proprietary accessibility API is out of the question.