The rule states that you shall not operate over people. If RTH is activated automatically due to loss of control, my interpretation is that you are no longer "operating" the aircraft; it is operating autonomously. Further, it is operating in the best interests of safety; returning to the point of origin to (presumably) safe conditions.
The FAA is almost entirely focused on safety in the NAS. You should plan your flight so that RTH wouldn't overfly people intentionally, but you can't prevent everything that someone might do. If you were to have an incident, the FAA is going to ask, "How did you address this risk?". You want to have a good answer; perhaps something like this:
"The flight path was designed to avoid overflight of people and was found to be clear of individuals both before and during controlled flight, however, during the return-to-home sequence caused by a failure of the control link, an individual in an area not part of the planned flight path but in the direct-to-home path chosen by the aircraft was inadvertently overflown."
Emergencies happen, and the FAA knows this. If you were able to fly in an area entirely closed off to non-participants and eliminate the possibility of anyone being able to get under the aircraft at all, no matter what, sure - that would be great. But real life doesn't happen that way.
What did you do to MINIMIZE - not ELIMINATE risk?
