A thought experiment. Not the kidney one, a different (though similar) one. A woman is newly pregnant against her will; she doesn't approve of abortion and isn't going to have one. She discovers the fetus has a very rare disease which is quickly fatal unless the fetus can be removed and implanted in a compatible host; such hosts are very rare but can be found via a computer search of a medical database. A compatible host is found. Is it murder if she refuses to be an actual host? Not just that - would anyone even think she had a very strong duty to be a host? Would anyone even think she had a weak duty?
I say no. Hardly anyone would think that. (Perhaps I'm underestimating the obsession with the fetus.) So the difference must be that in the usual case, the fetus exists because its mother had sex with a man. Why is that a kind of difference that makes a difference?
This is, as Benson notes, essentially a variant Thomson-violinist case; and the problem with Thomson-violinist cases is that we do often regard it as a violation of duty to refuse to allow temporary impositions on one's freedom when someone else's welfare is at stake. (It is another question, of course, whether it is reasonable to make such violations of duty illegal.) That's what duties usually are, in fact. And I think it is pretty clear that there are quite a few people would say that she would, in fact, have at least a defeasible and prima facie duty; and on such a view there is no fundamental difference in the two cases, although they may differ at times as to potentially exculpating circumstances. So I think Benson definitely is "underestimating the obsession with the fetus"; since the fetus is going to be regarded as a person by most pro-lifers, then the fact that the commitments involved in bearing it to term are unwanted is not going to carry any more weight than in any other case in which the dependency of persons forces us into unwanted commitments (e.g., fathers forced to support their children financially when they don't want to do so). From that perspective, there is no ground at this point on which to treat the cases as different, and from that perspective the bare fact of its being unwanted is not even relevant to the question.
What is still relevant on such a view, of course, is whether the circumstances involved in its being unwanted exculpate, excuse, or mitigate in any way. This is, I would suggest, why some pro-life advocates tend to be more sympathetic to Thomson-violinist cases than other pro-choice arguments: they can be read as scenarios in which circumstances introduce features, or are alleged to introduce features, that excuse in part or in whole, and thus in that sense make what would otherwise be 'erogatory' supererogatory. Whereas pro-choice advocates, I take it, tend to read them in the way Benson does here, as saying something about duties themselves rather than about how violations of such duties may be excusable under extraordinary circumstances.
To put it in more technical terms: these scenarios are casuistry, in the old respectable sense; pro-choicers tend to be laxist or probabilist about them, whereas pro-lifers tend not to be. There's considerable diversity among pro-lifers on the subject. Some, and fairly consistently the Catholic Church in particular, will deny that casuistry sheds any light on the subject, since a person's life is in question; others will accept the conscience-cases as genuine, but will tend be probabiliorist or rigorist in judging them. This is the fundamental problem with appeal to intuitions in a debate about this; moral intuitions are casuistical judgments, and while you can sometimes come to the same conclusions from different casuistical standpoints, it is your standpoint that governs how your intuitions will fall on this subject, and that is determined by prior commitments about (e.g.) the seriousness of the subject, what makes it serious, etc.