I will not go into more detail on Benatar's asymmetry. Benatar himself acknowledges that many people, upon understanding the asymmetry and its consequences (coming into existence is always a harm), are willing to claim that they do not see the asymmetry. Also, many have treated Benatar's conclusion as a reductio ad absurdum of his entire argument. In my own experience, I have heard it criticized as "more clever than deep" and too focused on the negative value of suffering, as opposed to other values - though, to be sure, always by those Benatar would describe as "cheery." As a (currently non-practicing) suicide, Benatar's arguments seem merely obvious to me.
For those who would let go of the asymmetry, or feel that the antinatalist conclusion is a sort of reductio of its supporting arguments, I feel there is a more palatable route into antinatalism from a rights perspective. Of course, there are many routes to antinatalism from a misanthropic perspective; I see human suffering as so particularly harmful that I am not particularly persuaded by them, but at any rate, this argument is a philanthropic argument, as is Benatar's.
Let us consider cases where one person inflicts harm on another without the victim's consent, where consent is impossible. Benatar draws a distinction (from Seana Shiffrin's "Wrongful Life, Procreative Responsibility, and the Significance of Harm," in Legal Theory, 5 (1999) 117-48) bewteen, on the one hand, causing harm without consent in order to prevent a greater harm, and on the other hand, causing harm without consent in order to provide a pure benefit:
Thus, we take it to be acceptable to break an unconscious (non-consenting) person's arm in order to prevent a greater harm, such as death, to that person. . . . However, we would condemn breaking that person's arm in order to secure some greater benefit, such as 'supernormal memory, as useful store of encyclopedic knowledge, twenty IQ points worth of extra intellectual ability, or the ability to consume immoderate amounts of alcohol or fat without side effects' (quotations from Shiffrin by Benatar).
However, Shiffrin and Benatar's intuition does not seem to be universally shared. Many argue that it is, in fact, completely acceptable to cause someone a harm in order to provide him with a benefit. While many harms parents do to children without their consent are in the interest of preventing greater harm (vaccination), plenty of other harms parents inflict on their children, with the approbation of society, are mostly or purely to provide a benefit, such as education, discipline, and indoctrination into a religion. Many people intuitively accept it as morally fine to strike or otherwise discipline one's child, or to force a child to study something he or she hates, or to teach frightening religious ideas to a child, in hopes that the child will thereby have a better life, one more in accord with the values that the parent feels the child will hold.
This is the point at which I interject myself. Why is female genital mutilation, performed on children, awful? It is awful because it causes physical suffering, and limits a girl's ability to enjoy sexual pleasure, perhaps. But if an adult woman chooses to undergo this body modification, in circumstances that lead us to believe that her consent is one hundred percent valid, we might reach the conclusions that Sheldon and Wilkinson reached in their article, "Female genital mutilation and cosmetic surgery: regulating non-therapeutic body modification," in Bioethics (1998 Oct;12(4):263-85); that is, that as long as genital mutilation is freely chosen by an adult aware of the risks, it should be allowed. So perhaps the harm of the genital cutting of children is a lack of consent.
Consent is the key to a rights-based ethical system. Why, then, should we allow a parent to consent to harm such as vaccination, teeth cleaning, surgery, and education of children, but not non-therapeutic genital cutting of those same children? All might be defined as harm to prevent a greater harm, or harm in the child's best interests, from the perspective of the parent. There are many ways in which one might try to distinguish genital cutting (it primarily serves the interests of those other than the child, it is a major invasion, its benefits are dubious when considered from a perspective outside the child's kin group), but none of these distinguish genital cutting from the procreation case.
Our legal system recognizes the principle, and I think it is a good one, that even a benefit must be consented to. A gift is not legally valid unless the recipient consents to accept it. Another problem for the harm/benefit dichotomy is that the harm/benefit distinction is often much less clear in practice than in the examples above. Why should it be morally acceptable to harm someone either in order to prevent greater harm or in order to provide a benefit? Both must be suspect in light of the bias that necessarily accompanies an agent's judgment of what is good for another.
I propose a general principle: it is ethical to inflict harm without consent only where it advances the values of the victim. (And the greater the imposition, the surer the perpetrator must be that the imposition advances the values of the victim.) However, knowing whether an intervention advances the values of the victim is extremely difficult in the absence of consent. If a person voluntarily consents to a harm, there can be little ethical problem with it (though there may be problems with knowing whether consent is truly voluntary, as with prostitution or other forms of paid work when resources are initially distributed unequally). Consent transforms rape into consensual sex, battery into medical assistance, slavery into employment, a forced march into a backpacking trip. We are free to take suffering onto ourselves. We are not free to impose suffering on others for our own ends, without their consent. This principle should give us pause and make us less sure about our intuition in thinking about even the "easy" cases, like vaccination, care of unconscious people (particularly attempted suicides), discipline, and education.
The problem with the birth cases - and, arguably, many education and imposition of religion cases - is that harm is done not to advance the values of the victim, but rather, to advance the values of the perpetrator (parent). Where we could accurately predict the future values of the victim, and had a good indicator as to whether the victim's currently expressed values should be ignored (attaining the age of majority seems to be an extremely poor hash), there would be little ethical problem with birth and education: children may be harmed to the extent that their future selves, as accurately predicted by our tools, would want them to be. However, accurate prediction is, of course, impossible (nor is it clear why future selves should take precedence over present selves, and exactly which future self we are here discussing). In fact, the predictions about future values are likely to be biased in predictable ways - optimistic, self-serving, and projecting the mind of the perpetrator onto the victim. A Baptist parent will assume that his child will wish to be taught frightening theology to be safe from Satan; nevermind if the child, once allowed to be free from harmful interventions, espouses Buddhism. I feel that a recognition of this principle is, at some level, responsible for the much less severe methods of education in use today compared to a generation before. We are queasy in the face of harming a child, even if we believe it to be "for his own good." For we are poor indeed at predicting the future good of another.
The ultimate unconsented intervention - a harm, which, of course, might turn into a benefit if consent could only be obtained - is that of being brought into life. Being brought into existence is a more serious intervention than sex, employment, or even bodily integrity, and yet we require no consent to be born. Of course, this is because it is impossible - there is no one to consent in advance of being brought into existence. But we should realize that this impossibility of consent does not excuse us, any more than the impossibility of consent of children to be genitally mutilated excuses their mutilators.
People have children to advance their own values, not those of the children they bring into existence. Put another way, procreators are using others (their children) as means to serve their own ends, without their consent. This is why it is wrong to bring new people into existence. An intervention as serious and potentially harmful as being brought into this world must be consented to; since this is impossible, it is wrong.
Note: the vaccination case, at least, is made slightly more complicated, but more in line with intuition, when we consider that failing to be vaccinated imposes a potential harm on others in society. However, so does failing to get a flu shot, or failing to receive other vaccinations as an adult, which is not, so far, compulsory. Note also that I am certainly not one of those who believe that vaccines cause autism.