In a recent LiveScience article the following statement was made:
"And while metacognition can involve self-awareness, the "I" part of the equation isn't a necessary ingredient, Smith said. Scientists are not sure if other animals possess self-awareness."
So, once again while many may see this as a semantic issue, this muddling of definitions creates far too much unnecessary confusion when making such determinations.  With respect to self-awareness in animals, I want to distinguish it from the sense of consciousness, or conscious awareness, with which it is often associated.  Specifically, the definition I intend to use, is to identify any aspect of an organism's existence that indicates that it has be "aware" of itself as distinct from anything else in its environment.  In addition, another characteristic of self-awareness is that it must have some sense of an identity with which it considers itself separate from others around it (i.e. its peers).

From this definition we will consider that conscious awareness, or consciousness is when, as in humans, we are aware that we are aware.

The reason why this definition becomes important is that, in the absence of self-awareness, we are left with the problem of animals (including microorganisms) having to engage in activities for their own survival, but suggesting that they are completely incapable of recognizing anything that would help that achieve that end.  In short, there can be no "survival" if an organism doesn't even recognize that it has an existence. Just to reiterate, this doesn't mean that an organism must be capable of "knowing" that it exists, but only that it possess the mechanisms to recognize its existence among others and its environment. The key element of self-awareness is the ability to discriminate among a range of choices.

If we take this definition, then we need to be able to identify whether other organisms are capable of recognizing themselves as separate from their surroundings and other organisms, and how such a thing might be manifest. Such awareness might be quite different from what we normally think of, ranging from smells or visual cues to recognition of molecular "signatures", but as long as it provides a means of recognition then it is sufficient to establish that an organism is aware of itself.

Given the full range of organisms from microbes to mammals, and regardless of how sophisticated or primitive the actions of an organism may be, they are all marked by a singular capability of engaging in activities to advance their own survival.  

From the simple premise of knowing that food should not consist of your own kind, one must conclude that all living organisms are capable of recognizing the group to which they belong.  This provides a reasonable basis for concluding that all living organisms must be capable of seeing that they are individuals belonging to a greater whole.  In short, even the most rudimentary form of self-awareness is a necessary condition to avoid cannibalism (1) and to achieve reproduction.

An example of this occurs with bacteria (Paenibacillus dendritiformis) where, when subjected to stressors, will release a chemical which kills off part of the colony, so that others may survive. This is not an unregulated activity, so the entire colony isn't wiped out, but only a portion of it.  It is important to recognize that there is no claim that this is a thoughtful directed activity, but rather that it only makes sense within the context of an organism that must be aware of its surroundings, other organisms like itself, and the fact that it is an unusual action used only during periods of extreme stress. 

We also can't argue that a purely mechanistic explanation makes self-awareness different in "lower" organisms because even human experience could be described in a similar fashion.  The fact that I may recognize a friend because a photon interacts with the retina sending a electro-chemical signal to the brain where neurological functions retrieve the memory doesn't change the fact of what is being experienced.  Neither does it change the nature of recognition if it is simply registering a chemical "signature" that distinguishes an organism from others.

We also know that many cells recognize each other by glycoproteins and will react negatively when intruders are present (infections or foreign tissue as in transplants).  Once again, in considering bacteria, it is clear that they possess a rudimentary social system whereby "awareness" is a necessary condition for communication.

While this may seem a matter of degree, it isn't substantially different from ants of the same species recognizing that they belong to different colonies.
“We are not pathogens for each other, so why when you get a liver transplant from a genetically mismatched person do you reject that liver?" 
In truth, animals groups are not pathogenic to each other either and yet they are quite specific and particular about recognizing members of their own social group versus intruders

All in all, the ability to recognize members of a particular biological group is a fundamental aspect of survival, and the basis for the cooperation necessary in larger multicellular organisms. It appears that the concepts of cooperation and society may have far deeper roots than anyone's ever suspected, extending all the way down to the cellular level and imparting a degree of unity in related organisms.

In the end it may be this simple principle of self-awareness that distinguishes the living from the nonliving.
(1) While cannibalism certainly exists, the point here is that it is not considered normal food-gathering behavior for any species.  Even if we consider the sexual cannibalism of insects, there is evidence to suggest that this is specifically related to providing nutritional value to an impregnated female and in controlling sex ratios.