It seems to me most social platforms (not just big tech, smaller UGC sections like the BBC) have many puppet accounts that are triggered by certain content.
Anecdotally looking at BBC comment sections of Scottish content, the "highest rated" comments are almost unilaterally pro-British/anti Scottish National Party which deviates a long way from historical voting preferences. The SNP have performed very well in Scottish and Westminster elections and the weakest barometer for them is/was the 45%/55% vote split in the Scottish independence referendum 12 years ago. I think if anyone took a "sentiment score" of what's there vs how people generally think or behave there'd be a large deviance.
More generally, any platform seems to have systemised abuse and this pattern goes all the way back to generic content management systems being abused in the early 2000s.
I do wonder, are these accounts being accessed via proxy? i.e. someone claiming to be from the UK and having a residential IP- if the platform doesn't care about the location of access, maybe start checking for latency?