However, increased consumption among youths leads to negative externalities. Raising the share of adults registered as medical marijuana patients by one percentage point increases the prevalence of recent marijuana use among adolescents and young adults by 5-6% and generates negative externalities in the form of increased traffic fatalities (7%) and alcohol poisoning deaths (4%).
In case you think there is no connection between the first sentence and the second, there is. In the paper for which this is the abstract, she does claim that people who use marijuana and then cause traffic accidents and/or die from alcohol poisoning are creating a negative externality.
Let's consider the two alleged externalities in turn.
First, traffic accidents and fatalities. There certainly is a strong case to be made that people who use marijuana and then cause traffic accidents and fatalities are creating a negative externality. But the case must be made that the accidents hurt other people, damage other people's property, or kill other people. She does not do that. Moreover, she does not appear to put any weight on the distinction between killing oneself and killing others. If I, through some foolish act, kill myself but don't damage anyone else's property or injure or kill anyone else, I have not created a negative externality.
Second, alcohol poisoning. Here the reasoning is even more straightforward. If using marijuana causes me to die from alcohol poisoning, presumably because marijuana and alcohol are complements, I have not created a negative externality. I have done something destructive to myself, but there is no negative externality.
I looked elsewhere in Ms. Smart's paper for some hint that she was aware that she needed to make a case for her externality claim. This, on page 18, is the closest she came:
If individuals are rational and fully anticipate the potential negative consequences of marijuana consumption on future utility, then any increase in use induced by medical marijuana availability increases consumer welfare (Becker and Murphy, 1988). If, however, individuals underestimate the potential negative consequences (e.g., risk of dependence) or make mistakes in their consumption choices triggered by environmental cues, this increased marijuana use may decrease social welfare (Laibson, 1997; Bernheim and Rangel, 2004). Consumption by children is often assumed to be suboptimal prima facie, as reflected by the range of policies designed to guide their consumption decisions (e.g., the minimum legal drinking age, compulsory schooling laws, and debt contract age limits). Under any theory of individual decision-making, if marijuana use generates negative externalities, then the socially optimal level of consumption is below the individually optimal level of consumption achieved under a free market regime.
In other words, she jumps from entirely plausible claim that young people may make bad decisions that are not in their long-run self interest to the non sequitur that those decisions create negative externalities. But recall that a negative externality is a cost imposed on someone else. If I kill myself, I kill my self.