Terminology note
We should not perpetuate the propaganda efforts of the commercial and ideological opponents of vaping by using their wording. This method of skewing the debate is primarily seen in their successful attempt to equate smoking with tobacco, and their current attempt to do the same with smoking and nicotine.
What they do is lie prodigiously and repetitively, and eventually everyone (a) believes the lies and (b) starts to repeat their lies for them. It's a win-win...
Smoking vs tobacco
Smoking is reputed to kill 440,000 a year in the US. I'm not going to get into that as I'm not an expert on the epidemiology. It's either right, or an exaggeration - but let's say it's correct, for the sake of argument. Let's agree, for the purpose of this particular position, that smoking kills, and a lot. Given that anti-smokers lie about everything, we cannot automatically assume their stats are right.
Smoking kills, tobacco use doesn't. There is such a huge amount of factual and evidential data from Sweden that a Snus consumer is equivalent to a non-smoker in terms of health outcomes, it cannot be argued. (That is to say, by statistical analysis, if smokers totally quit or switch to Snus, there is no reliably-quantifiable difference in health outcomes.) Individuals are as always exempt from the average because their genetics make them so.
Tobacco users in Sweden have proven (not just supplied some evidence) that tobacco use has no significant health consequences. In fact the exact difference can be measured, and has been, multiple times: long-term Snus consumption causes an average lifespan reduction of between 2 and 10 weeks. A 6 week average lifespan reduction is about the median, from the many large-scale multi-decade studies that have been carried out. A 6-week lifespan reduction is almost certainly less than that attributable to an equivalent consumption volume of coffee, for example.
What they do is to say "tobacco" when the reality is that it's smoking doing the harm. Therefore when discussing the mortality or morbidity resulting from smoking, we should correctly describe it as 'smoking-related' not 'tobacco-related'. It is playing into the hands of the propagandists to say tobacco-related mortality (death rate) when what you actually mean is smoking-related mortality. Tobacco doesn't kill, smoking it does. The same thing applies to tea.
Addiction vs dependence
Modern usage of these terms has become specific. That is to say, where perhaps in the past they meant the same thing, for those in professions that work with the consequences, there is a growing difference in the exact meaning. Another related term is reinforcement.
Reinforcement = the potential of a material or activity to cause repeated consumption or repetition that may lead to dependence. Reinforcement is a critical though not necessarily vital stage of dependence.
Dependence = a need to keep consuming a material or performing an activity that is abnormal by average measurement, but that has no measurable elevation of risk. It may elevate risk but this is unmeasurable in terms of modern urban lifestyles. In other words, it may or may not do harm, but it is difficult or impossible to isolate that from the background noise in modern life (or no one has done so as yet). Coffee is currently seen as a dependence, where applicable; though it looks as if it will soon become fashionable to investigate harm caused by it, if any can be measured.
Addiction = a dependence that has significant risk of harm. Harm may be of many types including physical (and thus to health), economic or social, and in some cases the harms may by implication involve others. So an addiction can be to a harmful drug like her_oin, a harmful drug cocktail and/or delivery system like smoking, a gambling habit (due to the economic harm to the person and/or family); and the social harms inherent in the activity (smokers 'are lepers', gamblers are 'dangerous to know').
So when we discuss nicotine, the most that it can be ascribed in modern usage is a potential for dependence. Addiction is incorrect since no one has ever demonstrated that nicotine (not smoking) has measurable harms. The Snus and NRT data are very clear on this: nicotine consumption has no clinical significance, and in fact they have no statistical significance either (a smaller measure of impact).
Smoking is an addiction, nicotine consumption is a dependence. Nicotine dependence is created by smoking (and possibly by tobacco use other than smoking - this is unclear at this time), but there is certainly no good evidence base showing that it can be created except by tobacco consumption.
There is an interesting question regarding vaping: strictly speaking it is neither, since no one has either demonstrated it to be reinforcing among never-smokers (of course, we must use those unexposed to tobacco for this test), nor to have any risk elevation that might impact health. Some might perhaps assume the worst and regard it as having the potential for a statistically-measurable impact on health outcomes (around a 2% repeatedly-demonstrated effect); although the 'vectors' are not visible at this time (no materials in vapor can be ascribed any toxic potential that would significantly impact health - see Burstyn/Drexel etc.). The likelihood of it having any clinical significance (killing or causing serious disease in more than 3% of participants) seems rather small.
Thus, the pedantically-careful would be correct in describing vaping as a dependence, if only someone could reliably demonstrate it as producing clinically-significant levels of dependence in never-smokers. This also seems rather unlikely, given that there isn't a single clinical trial that reports pure nicotine can create dependence in even a single subject unexposed to tobacco, but multiple clinical trials where large quantities of nicotine were administered to never-smokers (numbering, in total, in the hundreds) for up to 6 months and not a single case of reinforcement never mind dependence was noted.
It is entirely likely that two or three people, somewhere, may become dependent on vaping when they have never been exposed to tobacco first (which certainly does create dependence on nicotine, but that's another story: a story of 9,600 other ingredients, potentiators and synergens). A few out of 10 or 20 million doesn't add up to much. Somehow it doesn't seem likely that vaping will shown as having clinical dependency issues.
I'm not even going to add a disclaimer to this, as the facts are the facts.