The bots are really mad about Obama and Katrina, folks

Pulled together and made more narrative-like from a tweetstorm ….

People of earth, we need to talk about this Obama/Katrina nonsense. And it is nonsense, but not for the reasons you might naturally think it is.

First, let’s talk about the reporting. Scanning Twitter and pulling up a random selection of tweets from a pretty thin pool to somehow convey that the Internet is going crazy about Obama mishandling Katrina? Pretty lazy.

I get it. What idiots! Let’s pile on! And of course it is clickbait gold for publications desperate for revenue and focused on short-term thinking at the expense of long-term credibility.

But there’s a contextual problem here. The latest Pew research says Twitter has 330 million or so active users. Only 24% of online Americans use Twitter, which is about 21% of all Americans. Do the math, that’s about 60 million Americans give or take a few million.

Let’s say you, the intrepid reporter, find 100 tweets from real people (I’ll get to that in a sec). How can you credibly report that if you don’t contextualize it? I don’t care, make it 1000. Hell, make it 10,000. Same problem when something like 60 million Americans have a Twitter account. If the audience has no frame of reference, how can it judge how bad this problem is?

Of course, the frame of reference is missing for a reason. The point isn’t to expose, it’s to mock. So, mission accomplished. Bathe in the delight of your clickbait mockpiece that picked some random people and put them in the Twitter stocks for a few days.

But you are just one reporter. Your mockpiece went viral, high five! Now everyone is picking it up and we’re faced with this nonsense where Snopes has to commission some poor soul with writing a debunk piece.

(Humblebrag disclosure: My first assignment as a pro reporter was to cover a junior high basketball game. I got off better than the Snopes writer)

Bad enough as it is, but then there’s the bot problem (hi POTUS Putin!). Do a casual search of Twitter for Obama + Katrina prior to this news going viral. A lot of the accounts tweeting, even those cited by media, have markers of bots. Egg avatars, barebones profiles, long number strings in handles, weird letter combos in handles, lots of followers with a high following/follower ratio, etc.

Then look at the tweets themselves. They’re all the same information, structurally different, but the same stuff. These bots are sophisticated in their language structure. That’s one of the reasons they’ve fooled so many. It’s easy to spot a fake if 1000 bots are tweeting the exact same thing, but this is much more thought-out programming. If you look at the highlight details in most of the early tweets (Obama, golfing, Katrina, Trump helped) the details are all the same.

This guide has good tips for spotting fakes. Once you get good at it, it’s like seeing the Matrix on screen.

Anyhow, what we ended up with was what my brilliant colleague Bob Britten called a “cocktail” of bots, random idiots, and pranksters. And yes, there are actual idiots who thought to do this on their own. Every society has them. Let’s not lose sight of that. In a given population, social science tells us there are fools who believe dumb things. The fluoridated water, the staged moon landing, chemtrails, Roswell …. wait, Roswell happened. Anyhow, 1000 real people believing this isn’t really news, but whatever.

But while some were tweeting this from their confused brains, others might have just been parroting what they saw on pro-Trump hashtags. They’re downhill from idiot+bot activity, and it might as well be Fox News or Infowars to them. Let’s call them the second wave, the people who see the conversation and jump in without verifying a thing. Then there’s the third wave, all those bots circling back and either retweeting or copy/pasting tweets from the random idiots and pranksters. It’s like a circle of life except life is terrible.

But here’s my point: we’re being played, folks, and not just on Katrina. Read about the #FireMcMaster firestorm a couple weeks ago. This is discourse espionage, a long game by foreign agents to disrupt our conversation and news flow, and they’re being aided by people here at home who don’t understand how much of our discourse on social networks is being hijacked by bots right now. One estimate says 20% of Presidental-election-related tweets came from bots. That’s, uh, a lot.

What’s the goal? Who knows? Maybe if they can convince Americans to believe without context that there are many, many idiots in the population, it’ll be hard to get us to coalesce around an effort to stop real threats. Maybe they’re 4Channers at heart doing it for the lulz. Maybe they just don’t like us.

The sinister thing I see is that our discourse is being hacked. Democratic participation depends on us being able to talk to and argue with each other. But what kind of playing field do we have if we can be so easily tricked into thinking Americans so easily believe lies like the Obama/Katrina thing? Most of us don’t know anything but what we can perceive, and without knowledge of Twitter bot actvity or contextual numbers of tweets vs. population, it becomes very easy from a psychological standpoint to think you’re the lone sane person left on planet Earth.

But whatever the case, we need to be aware of how we are enabling this. Not just the retweeters and repurposers sitting in their living room yelling at Twitter, but also journalists who don’t understand the context here and aren’t smart about bot activity. Because it’s the news industry that is often unwittingly pouring accelerant on the fire.

I said before lack of context is a problem in terms of demographics (knowing how a few posts compares to total active users in a population), but journalists have to toughen their skepticism on bot trolls too. It should be required training in newsrooms that report on conversation.

Bots are here to stay until Twitter decides to do something about them.

What the news coverage did on Obama/Katrina was hand these troll bots and random humans a microphone and a platform. And now we are all amplifying it by sharing our disgust and linking to the stories.

So, to recap: A random search on Twitter yields some tweets written by bots and a few humans. The humans watching the pro-Trump feeds amplify it. Pretty soon it catches a journalist’s eye and they write about it, completely unaware of the demographic or bot contexts. More news sites pick it up. Pretty soon the fact-checkers have to do something about it because people are sharing it on social media. Meanwhile the whole story is getting noticed and spread by average folks despite being built at the top of Bullshit Mountain, with apologies to Jon.

I get it. Smaller, sketchier websites are still going to write about this. It’s nichey, outrage-oriented clickbait in a world of sameness, and people are not always good actors.

But pros are all better than this, I know you all are because I worked with you and respect your methods. You just lack the training and (let’s be honest) time to sharpen things up a bit.

So, my advice:

1. Newsrooms need to educate reporters and editors to spot bots in the wild. It will give them a useful way to think about whether to pick up a story that goes viral. And by the way, maybe the pickup if there is high bot activity is to point that that the viral story is a garbage story built on bots and a few random Twitter users. There are ways to add value to a story stream without going full Snopes.

2. Journalists and news consumers alike should exercise extreme skepticism about conversation stories that lack context about who is vs. isn’t saying something. I don’t expect consumers to do the sleuthing — that is the reporter’s job. But if a consumer sees a story that appears to have no sleuthing attached to it, that’s reason to be skeptical.

3. Consider your motivation for sharing. If it’s to point out how dumb random people are, consider whether you’re feeding the beast by passing it on. Outrageous things said by people with power are a different thing, of course, but publicly shaming a few random Twitter eggs only benefits the clickbait industry built around it.

4. At some point, and this is important, the real value on social networks is going to be in verification, how to represent real identity on networks so there is some sense that the people present are real, accounted-for, and accountable. The major social networks have to figure this out, or someone is going to do it for them. They depend too much on information sharing to keep people coming back and sharing, and people won’t use a product that is untrustworthy or full of fakes. The next wave in social networks is going to involve verification in some way or another.