Absence of evidence is not evidence of absence.No I'm perfectly aware. There's been no serious evidence that bot farms of any kind, from cheapo shell companies to state sponsored programs, have had any significant sway on an election.
I accept this is likely to be very difficult to measure. However, there's research going on which suggests well targetted bot campaigns and the like can be capable of shifting substantial opinion. Some of this goes outside politics and into things like conventional product advertising. There is surely a difference between voting and sales, but I think we would be extraordinarily complacent to think they are sufficiently different that the principles don't carry across at all.
Complacency is I think a big issue here. I am well aware that at least one study suggested that attempted Russian influence in the 2016 election had little effect. But that cannot be assumed as a proxy for all such campaigns: it might just mean that the Russian interference campaign in 2016 was poorly orchestrated.