I have a feeling that any framework for mitigating AI security/injection attacks should not involve AI as its core working as the post suggests. ACLs and separ

I have a feeling that any framework for mitigating AI security/injection attacks should not involve AI as its core working as the post suggests. ACLs and separation of "executable" instructions vs user-input data should be the way. Otherwise it's just another form of circumventing "injection protection" or "buffer overflow"-esque attack (where the user maliciously "overwrites" the supposed-to-be-hardcoded instructions that should be rock solid to the system, with maliciously crafted user input) will be discovered one after the other.