I believe "curbing" is simply a nod to the fact that the people/groups controlling these bots have no interest in turning them off, and are unlikely to change their opinion. The harms this article shows are, after all, their purpose.

The actors who may be motivated to mitigate these harms, i. e. the social networks and possibly governments would face problems very similar to those with (other) spam, namely identification. Of course the history of spam filters makes me optimistic that it's feasible put a rather significant dent in the problem with even just modest efforts.

Summary: It's more of a semantic difference, arguably more precise because there is no off switch as such.