Should Large Language Model inference or other processing by the user be supported?

There are circumstances in which the user may benefit greatly from processing aspects of the document (e.g., text or images) with tools available in the operating system or assistive technology. The reading application should allow this to be done, unless there are underlying reasons of intellectual property or social context (e.g., an educational assessment scenario) in which this functionality should not be supported.

Examples include:

* Using a Large Language model to describe or to answer questions about an image, beyond the text that the author has provided in the document.
* Using a large language model (LLM) to generate summaries of content - potentially of cognitive benefit.
* Transferring an image to a tactile graphics-capable display, such as a pin-matrix device.

I suppose this is largely an issue of access control, but there may be other support that the reading application needs to include.

No files, tests, or entry points are identified in the issue. Start by defining the reading application's requirements for user processing of text and images, including access-control exceptions and assistive-technology integration; done means the supported capabilities and restrictions are documented clearly enough to guide implementation.

Written by the indexing model from the issue text.