Do you need to use a "prompt" to generate the audio? Then it is gen ai. Telling a thing to read out loud a line is not a prompt. Telling it to read it in a spooky voice is. And if you can do such voice styles in the first place by telling the system, that also is a huge hint that it might be gen ai. Maybe read the docs of whatever system you use to find out what kind of voice generation it uses.
But what you describe will not be what you described. A ghost using radio waves and grabbing the audio in those waves to communicate would not sound like a retro robot. It would sound like Bumblebee in the Transformer movies.
I think you mean voices like from a radio station that is not quite tuned in and fades in and out and has static noise and such. Maybe combined with how Bumblebee would "talk". There might be filters in audio manipulation apps to achieve such effects.