You tossed those letters to us without considering that in the wider world, we don't automatically follow the lingo. If by GAN you mean Gaussian additive noise, then yes, that's a well-used source of randomness, usually by using amplified noise diodes or other random processes. If by GAN you mean generative adversarial networks, that makes inadequate sense. That involves finite state machines (computers) that can only approximate randomness. And if by GAN you actually mean gallium nitride, that makes even less sense.

I am sorry for not being thorough with my question. I mean Generative Adversarial Networks.