Language behaviour is learned behaviour and conscious behaviour. You learned your first language at a very early age - you learned how to make, and to some extent to hear speech sounds at a very early age - so early that you almost certainly don’t remember learning it, but nonetheless you did learn it. You do not do it automatically, or unconsciously.
All this “reverse speech” stuff is garbage, pure and simple.
If you listen to speech reversed through a tape recorder - or any other source of gibberish within the normal sound range of human speech: the output of someone with glossolalia springs to mind - you are used to trying to interpret these sorts of sounds as speech, and if you concentrate hard enough, you can come up with some sort of interpretation of the noise. But whatever you come up with will depend entirely on you - your conscious or unconscious prejudices about what you expect to hear - and has nothing to do with the speaker’s intentions, or hidden desires, or whatever. The interpretation is supplied entirely by the listener. (And it will depend on things like the listener’s native language - you learn to make phonological distinctions, based on your native language, so speakers of different languages can “hear” quite different speech sounds within the same set of gibberish sounds. Run a search for, say, “categorial perception” to get more information).
It would be relatively simple to design an experiment to demosnstrate this - off the top of my head: play a chunk of backwards speech to two native speakers of different languages, don’t give them any information about where it originally came from, and ask them (in their native languages) to come up with an interpretation. They should both come up with something - in their native languages.