I have been studying data compression for a while. For educational purposes, after a lot of reading, I managed to create a software that performs encryption and authentication using AES256-GCM. However, I have now reached a point after which I cannot find a possible answer and this is confusing me.
My software allows the user to choose the "hardness" of the encryption, a value ranging from 0% to 100% (a scale). Nonetheless, I have not managed exactly how I could mingle the logic of "hardness" with the logic of AES. Yes, I have a lot of ideas about playing with the key, the IV, the resulting data and so on, but a point I have learned on this area is that you should never reinventing the well when there is already studies on the same aspects and whose results had been checked and rechecked over and over and over again.
Therefore, what would be the best way to perform such work? I seek for theory, not necessarily coding.