Bots Operation

We've been able to determine that our account is being targeted by some kind of bot farm, likely for malicious purposes. We also have drawn several inferences which we consider to have a high probability of being true. We are attempting to cleanse our user list using an automated tool, however its effectiveness is limited. Here we will explain what we know, our degree of certainty, and how we've reached our conclusions.

What We Know 100%
According to tools like Twitter Audit, there were approximately 50,000 bots among our followers, as of a few weeks ago. Some are relatively mundane bots, but many are highly sophisticated and have a very strong appearance of being an actual human being, to include holding rather complex two way conversations. In recent weeks we've noticed a high amount of activity from suspicious accounts, including many egg accounts, pornographic bots, bots with names in the Cyrillic alphabet, and some bots posting in Russian. Other suspicious accounts looked like they might be real people whose suspicious behavior appeared more like real trolling because the conversation seemed to complex and sophisticated to be a bot.

Near 100% Certainty
It turns out that many (most?) of these seeming trolls are in fact highly sophisticated bots. Despite appearing like real people because of things like complex two way conversations (yes, it appears that in some cases we actually conversed with a bot) these accounts were merely high level bots. We are now virtually certain that many of these bots aren't simply stumbling across us, we are being deliberately targeted by whomever controls them. Whomever has chosen to do this is up to no good.

Very High Certainty
1 - The reason we have been targeted is to discredit info we may provide, attempt to identify us somehow, and/or to generally harass us and anyone who might engage.

2 - This bot attack is somehow associated with Louise Mensch and the Daily Kos writer "TrumpResistance", with the two likely being the same person.

Speculation
Some of these bots could be related to the effort we spoke about before by Steve Bannon to determine our identity and/or to discredit in lieu of identifying us. We are aware that the efforts to identify us continue, though we have not heard much as of late. The most recent info we have is that it's become a back burner priority, but that there was something specific that was being planned. While a purely speculative suggestion, the hypothesis that Steve Bannon may have sought to employ bots against us would appear to be consistent with what we are now seeing, and with the fact that Russian sources are known to have used bots for propaganda purposes as part of their interference efforts.

Reasoning
Some weeks ago we began observing a sudden influx of bot interaction with our account. For example, we might receive a notification that we have 80 new followers, of the 20 thumbnail profile pictures displayed 18 might be eggs. We also noticed that there seemed to be identifiable families based the similar superficial characteristics of bots (i.e. things like random characters usernames vs usernames mixing words with randomly generated numbers or other characters). So while some bots might have reached us randomly it was clear that many were related to each other, acting in coordination. Many of these have had names spelled using the Cyrillic alphabet, some posted comments in Russian. We also came to distinguish other tell tale signs that seem to be common among bots.

As we began paying closer attention to these suspicious accounts we began to notice other accounts popping up that were also suspicious, some having Cyrillic names and/or tweeting in Russian, but having believable profile and cover images, and appearing too sophisticated in their speech to be bots. We began paying very close attention to these as best as we could because they remained a curiosity warranting explanation. The behavior we observed by these accounts could be summarized as generally poorly reasoned arguments against things we had said, or things said by people agreeing with something we said. These accounts appeared to engage in two way conversation at a higher level of complexity than we would have expected from a bot, and so we initially believed that these were just trolls and that the Russian speaking by some was simply a coincidence.

Nevertheless the frequency of these apparent live trolls appearing took a sudden uptick, and we noticed that while their two way conversation was seemingly too complex to be bots, there also seemed to be consistent maximum level of complexity. While they could reply to a reply in a way suggested understanding, third level replies were more problematic. We then began to notice similar reply behaviors from other accounts that did not show any use of Russian language or Cyrillic alphabet. As we began looking closer to these suspicious accounts we noticed some similarities to simple bots, such as suspiciously low number of followers, following an abnormally high number of accounts, etc. One of the most interesting observations we made was that many of these seemed to have few to no original tweets of their own. Predominantly they retweet what others have said (mostly from large and often verified accounts) or reply to replies to these accounts' tweets. And nearly all the conversations were political in nature.

Based on this we began to realize that it appears that we have indeed fallen into the cross hairs of sophisticated Russian based propaganda-style bots. And while surprising, on its own it is still more curious than indicative. Feasibly, we might have simply been followed by many of these bots due to the simple nature of our feed. We had no significant information that might imply intentional targeting. At least, not until today.

Today, March 23, at 2:33 pm eastern time, an individual named Josh White tagged us and Louise Mensch in a tweet, speculating that a hidden meaning within our website's code was a shout out to Neo-Nazis. The speculation is insanely far fetched and revolves around the fact that a certain coffee shop in Northern Virginia is housed in a building which, nearly 100 years ago, was used by American Nazis. (White, we assume, is simply engaging in idle, albeit absurd, speculation.) A short time after that tweet, at 3:07 pm eastern time, LM replied jubilantly, with an allegation that the FBI is investigating us. At 3:44 pm eastern time, mere 37 minutes after LM's reply, and 71 minutes after Josh White's tweet, a Daily Kos writer TrumpResistance published the -- er what is the correct word for something on Daily Kos seeing as article seems too dignified? Whatever it should be called, the Daily Kos link we shared earlier today, based on White's tweet.

That is when a very bizarre event occurred. Within minutes of the article going live on Daily Kos, we were blasted with more notifications than we could count, informing us that we had been mentioned in a tweet. Those mentions were, in fact, posts sharing the Daily Kos link. That is actually how the link came to our attention. That there was a sudden rush of notifications of seemingly hundreds of people sharing this link within moments of it going live was incredibly suspicious. And as we looked closer we saw that these notifications were coming from bots who were sharing this link just moments after it went live. We later shared the link ourselves.

And as we took a closer look at the link we noticed the author was TrumpResistance. That caught our attention because we were already aware of allegations that LM publishes on Daily Kos using this name. And further review of the link gives the appearance that it's more about praising LM than anything. The author, for no clear reason, mentions in the comments LM's early attempt to rationalize her allegations that we are controlled by Steve Bannon. In other comments the author alleges that LM is the target of dual organized conspiracies by the far left and far right on Twitter. We understand that this is a claim that LM has made herself in the past, and is clearly consistent with her tin foil helmet worldview, but it seems very difficult to see anyone but her making such a claim.

We're sure that everyone will remember how LM became obsessed with attacking us after we refused to reveal our IDs to her. She has previously sent minion-ish followers to spam on our feed and harass those who follow our account. Based on all of this information, it seems highly probable to us that LM is closely aligned with, if not in fact the same person as, the author of the Daily Kos link. The elapsed time between today's tweets and the publishing of the link aligns perfectly. The nearly instantaneous bot activity sharing the link was very clear and convincing for us that the author has some connection to at least some of the targeted bot activity we've observed. And it offers a new context for us for the assortment of suspiciously organized bot activities we began noticing and monitoring weeks ago.

The true significance that needs to be seen here is the far reaching power of automated propaganda in our modern world, how zealously it is being used to try influencing political discourse, the fact that this influence is predominantly foreign in origin, and especially the fact that this propaganda is aimed at discourse unfavorable to the President. Our American society has lost the ability to distinguish between tabloid rabble and intellectual news coverage. We need to claim that distinction back. It is not enough to refer to outlets like Breitbart, Daily Kos, Addicting Info, or InfoWars as biased news sources. We need to stop giving them the lofty status of being news sources at all. The invasion of tabloid fodder into the realm of news isn't a first amendment protected issue. To the contrary, pushing back against all of this emerging tabloid fodder is an affirmation of the importance a robust news industry in our democracy.