It's interesting, as the closest analogue to this sort of argument can be seen with how support for gun rights are often framed in America. In that argument, one has a tool that can provide positive outcomes (protection of self and others) and negative outcomes (homicide, suicide, manslaughter), and emphasis is placed on the user's role in any negative consequences. In America where this sort of thinking is widespread we have more guns, and thus have far more negative outcomes than the societies that more strictly regulate guns, and thus have fewer guns in general.
I expect we'll see a similar split in how AI is treated by societies, with more negative outcomes being seen by societies which regulate AI less.