In case of a complete power failure, a light-weight ballistic rescue parachute will let the whole thing glide to earth. There might be some altitude where the parachute doesn't have enough time to deploy, but that's really the same problem as with a helicopter (where autorotation is only possible from certain altitude/airspeed combinations) - so you have to avoid staying in that critical region.
I think it is premature to make pronouncements on the safety - piston engines don't really have a great safety record. Turbines are amazingly reliable of course, but the entire helicopter ensemble still fails occasionally.
Electric motors and batteries (plus maybe petrol powered generators, or fuel cells) are just an entirely different technology stack, whose reliability remains to be seen. I'm optimistic, though.
If all the drone controls are wired to one battery controlled by one computer, yeah, there's a lot of single-points-of-failure there, and the result is scarier than a helicopter to me.
On the other hand, if it was basically a whole bunch of independent drones with pretty much their own electronics (I'll permit for convenience a single charging bus), and communicating with active-acknowledgement protocols such that each drone assumes the others are failed unless there is active assurance they are on and working, the whole ensemble can come out much safer than a helicopter. In theory you could get to the point where it doesn't just have no single point of failure, but even no dual- or treble-points of failure either.
Depends on the details.
The 18-rotor Volocopter can have several rotors fail with no problem before the ballistic rescue parachute becomes necessary.