Theoretically, yes, it's impossible to detect this perfectly when the attacker can do whatever they want to your code and memory.

Practically though, it's exactly as we see here: a kind of arms race or cat and mouse game where the attackers circumvent the detection and the detection evades the circumvention.

So no, I don't think it's futile.

They're not detecting the thing malware does, they're detecting the thing ordinary users do when they want control over their phones. Malware is different code so it has a different signature -- purposefully, because it has a strong incentive to not have these naive jailbreak detection methods work against it.

So all they're doing is detecting false positives from innocent people and zero instances of real malware.

Why would that be impossible for malware with root privileges?

Impossible to detect.