Read on Twitter
Currently some divergence between various stakeholders investigating disinformation and social media manipulation. Part of the challenge is that different stakeholders see different pieces of the puzzle and have different approaches to filling in the missing pieces 1/8 https://twitter.com/yoyoel/status/1263944345637412864
Around significant geo-political events a range of actors mobilise to take advantage. There may be different actors and categories of activity in play but at times their objectives, and even their behaviours, may overlap 2/8 https://www.aspistrategist.org.au/chinese-diplomats-and-western-fringe-media-outlets-push-the-same-coronavirus-conspiracies/
The overlap may look so similar that different sets of activity might be categorised as the same phenomenon eg by machine learning or by researchers blurring covert personas with unwitting proxies 3/8 https://twitter.com/ngleicher/status/1264614996878622721?s=20
Malign actors seek to drive authentic engagement. May use or mobilise proxy networks to do so. May wish to drive sentiment as much as shape narrative, to create a permissive environment. Where in this spectrum of activity to draw the boundary between authentic/inauthentic? 4/8
This is a particularly tough gig for the platforms in making judgements on state-linked influence operations. There's much we can do in terms of the precision with which we describe the phenomena we investigate in relation to state actors 5/8
In our work @ASPI_ICPC we try to be very clear on the categories of activity that we investigate. Whether it is state-linked influence activity.... 6/8 https://www.aspi.org.au/report/tweeting-through-great-firewall
....or patriotic trolling that aligns with state interests 7/8 https://www.aspi.org.au/report/covid-19-disinformation
For the methods people....we use AI and ML to discern signal from noise but also the expertise of linguists, analysts and researchers who provide qualitative contextual insights that make sense of the analytics 8/8
