It is always very dangerous and misleading to argue about abstract categories as if they were real things outside of the minds that make them. "Addiction" the label we use for a category we impose on the objective reality we observe. The OP is questioning the utility of that categorization based on more extensive observation of that objective reality.
In particular: there is evidence that the population we now group into the category "addicts" may in fact include two sub-populations, one of which "ages out", one of which does not. We currently have a "lumped model" of addiction, which associates certain additional features with the phenomenology we use to assign something to the category.
For example, if we find a person who is a chronic user whose use is negatively impacting their lives and who seems unable to stop, we are going to label that person an "addict". But the members of that category also get assigned other attributes for free. In particular, the current model of "addiction is a disease" says that anyone who fits the external phenomenology is also never going to get better without "treatment", or at best is very unlikely to do so.
But that association of "uncontrolled harmful use" and "its a disease" is a purely theoretical construct that may be wrong.
For comparison: I once worked on the genetics of a particular type of cancer that had two very distinct outcomes but which presented identically in a clinical setting. A physician I worked with described it as being incredibly frustrating, because she could literally see two patients in the same day who were apparently in the same condition based on histology, stage of the disease, everything, but ten years later one would be watching his daughter graduating from high-school and the other would have been dead for nine years. Simply because the diseases looked the same by some measures did not mean they were the same thing.
So it could well be that the same external phenomenology is present in two quite different conditions. It would be extremely silly to argue that one is "really" addiction and other is not, because "addiction" is just an abstract category we created to subsume different instances of similar phenomenology. We could call one grue and the other bleen, for all the difference it would make.
The possibility of two quite different underlying conditions presenting similar symptoms is one that has been realized often enough to be worth taking seriously, and when it occurs in things like the cancer described above no one ever makes the argument 'well X sounds like it is really cancer type Y to me'. Yet when the question of addiction comes up, many people say precisely that, and I am at a loss to understand why.