It may seem like I spend an inordinate amount of time on this subject, but it came up when I was listening to a debate between noted atheist philosopher A.C. Grayling and the preeminent Christian apologist William Lane Craig on the podcast Unbelievable?, discussing whether belief in the Christian God was reasonable given the fact of natural disasters like tsunamis and other so-called "natural evil" (i.e. suffering which is not attributable to the moral agency of a human being). The debate covered the standard arguments, but I think Craig's talent as a debater and rhetorician overcame Grayling's talent as a philosopher and no solid answer was given to Craig's assertion that the naturalistic world-view can not offer a foundation for the existence of objective universal morality, the way that a supernatural, theistic world-view can offer such a foundation in the existence of a universal moral law giver, namely God. He illustrates this by creating a straw man naturalist who explains the existence of moral values as emergent characteristics of the evolution of the human species. This explanation clearly does not allow for objective universal morals because morals are culturally and evolutionarily determined and thus one cultures moral values cannot be said to be any better than another's. To illustrate the dramatic consequence of this view on morality the favorite example is that we cannot say that the holocaust was wrong, because one society's culturally and evolutionarily determined values are as good as another's. Nothing exists outside of culture and evolution which can say these moral rules are the ones which everyone should follow, so in a way acting immorally is equivalent to acting unfashionably. Furthermore, if we could rewind the evolution of the human species and play it forward under different circumstances, we could have developed very different moral values: rape might be okay, it might be considered abominable not to kill your children if they are born with blonde hair, et cetera whatever. In the several debates that I've heard where William Lane Craig raises this issue I have yet to hear any one give him a satisfactory rebuttal, but it seems to me that there is a very clear answer to this argument.
Reason is the foundation for objective universal morality in the naturalistic world view. Once you have defined morality as a system of guidelines in place to promote the well-being of sentient creatures, and to prevent their suffering, then reason does the rest of the work in telling you how to get there. No doubt Dr. Craig would object that I have snuck in an assumption to my definition by saying that morality has to do with the well-being of sentient creatures, and ask for the foundation of this definition. But I think that is an absurd demand, we all know that morality is dependant on the well-being of sentient creatures, as intuitively as we know that the field of medicine is about the physical (and/or psychological) well-being of sentient creatures. We don't ask doctors the philosophical foundations of their assumption that we want to be physically well before they treat us, we assume that they know as well as we do that engaging in the practice of medicine assumes a benefit to physical well being is desired by its participants. In the same way acting as moral agents we know that engaging in the practice of morality assumes that the desired outcome is well-being among sentient beings.
Notice that morality, in this iteration, is not founded in culture or evolution and is even wholly independent of culture and/or evolution. If I were able to construct a purely rational artificial intelligence, meticulously absent any emotion, cultural or evolutionary characteristics, and program it with this definition for morality, it would be able to act morally.
The only questions left are about which ethical system best accomplishes the goal of morality.