It turns out when you actually get tested for HIV, they do two tests. One has a high false positive rate but a low false negative rate. If that comes positive, they do a second test that's the other way around.
It also turns out that we don't test everyone for HIV. In what I am sure is some form of irony, we probably test more folks for carrying bombs than we do for having HIV. When the numbers show there are more folks with HIV in America than there are those carrying bombs.
What that means is that people who trip these explosives detectors should be politely treated while it is determined whether or not they actually have explosives, since they are likely not really carrying explosives.
But this doesn't mean that scanning for explosives is a bad idea, or that people should just be merrily flagged to go on their way, otherwise what is the point of doing any screening?
After all, the vast, vast majority of passengers are just trying to go from point A to point B so each and every single scanning method will generate thousands of false positives every year.
If you subject a random American to this HIV test and the test result is positive, what is the probability that that person has HIV? I wont do the math, but the right answer is about 2%. Since the prevalence of HIV is so low it doesn't matter how accurate the test it, it will still generate many times more false positives than true ones.
Same thing with bombs. The number of American air passengers/year is about 800 million/year and of them, at most 5 are carrying explosives. Even if the bomb detector is 99.9% accurate it will generate thousands of false positives for each real bomb it catches.