,
12 tweets,
5 min read
Read on Twitter
A brief thread on the need for rigor and responsibility in the tools, methods, analysis, and presentation of #OSINT research, particularly if being posted publicly, even more so on Twitter since giving full context is even harder than written pieces. 1/
OSINT tools can be a double edged sword: same tools I use to verify identities are regularly used for doxxing; the same trails of breadcrumbs could assist a stalker or other bad actor, not to mention their potential analytic value. Use responsibly. 2/
Methodology is also important; this is a science, so we need a null hypothesis (this account is totally who they say they are, and that’s their Twitter behavior). The focus is not on proving your theory, it’s finding enough evidence that you can confidently reject the null. 3/
By this I mean; have you ruled out external influences (news worthy events, post goes viral)? If they’ve been kind of enough to give you a name and location, have you looked for that person in that location? Checked to see if the username was used elsewhere? 4/
If they provide a location, mention a place, or post a pic, have you checked to see if that’s the right place, checked local news for events that could explain differences in behavior? We tend to have tunnel vision that everyone uses Twitter the same way. They don’t. 5/
Are you using the right tools for the analysis? I have a number of different tools for different tasks, a lot of which may be used in any individual project. But a lot of consideration goes into what type of analysis needs to be done before you start, and that’s a vital step. 6/
Finally, presentation. The disinfo community has in some ways been too effective for its own good. Someone’s disagreeing with you? The common response at the moment seems to be, “Russian bot”. Even well-respected news outlets are doing this. Recently one posted re the debates 7/
While bot’s in the headline, by the analysis they presented, even being favorable in the amount of bot-like accounts they identified, they’d account for 8.9% of the engagement being described. This doesn’t help the public in understanding the threats and ends up divisive 8/
We need to be better at talking to the public, paying careful attention to terminology. Is it a bot? Troll? Cyborg? Some puppet variation? These distinctions matter. As does how we proved are assertion, particularly if making an attribution to a specific actor or country. 9/
There are a lot of great people on here doing their best, largely volunteering their free time, to help fight this manipulation of our civil discourse. But flawed research spread as fact is damaging to all of these efforts, and makes bad actors jobs easier. 10/
End rant. A few people I’d recommend on the topic : @conspirator0 + @ZellaQuixote, @SlickRockWeb, @jasonbaumgartne, @3r1nG....I’m sure I’m forgetting some and will add as I think of them. FIN/
A few I missed the first go around (my apologies, it was late): @NetwarSystem @ushadrons @josh_emerson @d1gi and of course @benimmo
