They might be "true" random numbers, but hardware implementations could introduce some noise which might bias the distribution in such a way to give some advantage on exploiters (I'm assuming a cryptographic usage).
If I remember well, also a quantum encryption (unbreakable) scheme was broken exploiting "practical" issues like that.
More like "random enough." Quantum fluctuations are not necessarily random. We just don't have the technology to understand the forces and processes which influence their fluctuations. Randomness is a human concept I don't believe truly exists in nature.
As far as I know quantum physicists already proved that there can be no "forces and processes" which can make quantum physics effects predictable. Einstein didn't believe in that, but it was proved.