Part of the problem is that in psychiatry/psychology, and so also in this test, the technical definition of "empathy" doesn't match how the word is used in normal conversation. It's a much narrower sense of the word, along the lines of "do you instinctively have an idea how a certain situation is making some other person feel?", or "if you want to make another person feel a certain way, do you instinctively have an idea how to do that?" Whether or not it is instinctive or intuitive is just as important as whether you can usually work out the situation correctly. And it's always about your perception of another person's emotions, the fact that an event might make you have your own emotions that are very powerful doesn't enter into it.
It's possible to feel sympathy with someone in need, and so treat them considerately, even though you worked out how they are feeling in some other way. For example, maybe they just came out and told you, or you spent some time comparing the situation with memories and advice you've been told. But that's different to what the test is checking for.
The test is not intended to learn anything about a person's morality, the strength of their emotions, or how considerate they would be once they know that another person is in need. The trouble is, of course, that most people reading or hearing about "lack of empathy" will most likely take it to mean the everyday definition of the word, including a lot of healthcare workers (and many of them are confusing "empathy" with "sympathy" anyway!)
Whether any of that makes it a good screening test for autism is a whole different matter, of course. But whether it is a useful test or not, getting a low score doesn't make anyone a "monster" or a psychopath!
Just for the record, I scored 19 when I did the test at my formal assessment. (Prime number! )
_________________
When you are fighting an invisible monster, first throw a bucket of paint over it.