I realize it probably sounds like a good idea, but screening and training volunteers is likely to be even more expensive than keeping the minimal crew they have operating the rover now.
Ditto on the API and "simple prototyping program," although I'm not even sure what that second thing even it supposed to be. Just writing the software would be expensive, not to mention you now have to have NASA people reviewing what comes in.
Driving rovers probably sounds a lot easier than it is. Commands are strung together in sequences. Sequences have to be checked to make sure they don't have conflicting commands. Instruments have to be taken into account; it's not just driving around that is being done.
And then there is the intense analysis and investigation that has to be done if something goes wrong. Reports have to be written explaining everything down to the bit level.
On top of that there is planning to be done to even decide where to drive, which involves a whole lot of people.
There is not only no money to be saved by handing operations over to "the community," there is also the probability that if you did the rover would be ruined within a few days.