Difference between the actors in the existing (non-automated) and the new (automated) system [on hold]

I guess two main use-case diagrams in building an automated system (Software) from a non-automated system are:

1. Analysis of the existing system by a use-case diagram.


2. Mapping the current use-cases and requirements to the software use-case diagram.

How then the meaning of the actors differ in these two diagrams?

For example I want to implement a software for a non-automated library. Is the librarian an actor of the existing system or maybe part of the system itself? Does the librarian turns to an actor when I draw the use-case diagram of the software system?

I myself guess when I want to analyze the existing system, I should look it as a library with its all staff and services, then the only actor is member!? But when I want to build a software for it, I should suppose the software as the system and everyone who works with it as an actor. Are they correct assumptions to find the actors of the existing and new system?