A very cool article on rumor propagation and fact-checking in the twittersphere

The U.K. Guardian has this fascinating piece on the way “…misinformation corrects itself in open, unregulated forums. (The) initial source was a corpus of 2.6 million tweets provided by Twitter, all of which ‘related to’ the (London) riots by virtue of containing at least one of a series of hashtags.”

Once you’ve clicked through and watched the rumors run their course through the twittersphere, be sure to click on the link to see the thinking and process behind these interactive graphics. Because I don’t like all-text posts and because I want you to become intrigued enough to check out the story for yourself, here’s a static screenshot of the interactive.

When you hit the “play” button in the upper left, the slider moves along the timeline. The circles, each a cluster of retweets of a particular tweet, change in size dynamically over time. Click on any circle and you can see the actual tweet and who posted it. Green represents tweets confirming the rumor, yellow asking questions and red refuting the rumor. This (false) rumor had to do with army tanks being deployed in London.

As you watch these, remember that some of these timelines are a few days long, others are a few hours. It is an incredible example of crowd-sourced fact checking. I would still guess, however, that some of these rumors have persisted outside of Twitter to this day. Someone who saw one of the original (false) tweets about rioters breaking into the London Zoo and releasing animals, including a tiger, but who never saw the refutation hours later, probably still believes it happened. Likewise with those who heard it word of mouth, Twitter participation or not.

On the other hand, any rumor-believer will eventually repeat it to someone who knows the truth and can correct the misinformation. An excellent companion project to this would be a survey of the population to see how and whether these rumors, started and debunked on Twitter, have entrenched themselves in the public consciousness.

I remember all the skepticism that surrounded the launch of Wikipedia. How could such a thing ever be accurate? Well, according to a survey of research on exactly that topic on Wikipedia itself, it turns out to be quite accurate, despite the ability of anyone to edit virtually anything with complete anonymity.

Wikipedia, like Twitter, relies on users to self-correct bad information and enhance and propagate good information. Misinformation (or “vandalism” on Wikipedia) decrease the quality of experience and utility of the platforms and so are seen as damage. Adaptive networks deal with damage by routing around it, isolating it or repairing it. Networks like these have self-evolved curation processes, each one unique and adapted to the system that spawned it, but all working in similar fashion.

(Hey, if anyone has links to articles on this subject, send them my way, eh? I’ve run across articles in the past about how content curation evolves in networked communities, but couldn’t find any of it tonight.)

Anyone can post nonsense to Twitter and hashtag it however they please, but the community uses its own evolved (and evolving) feedback structure to isolate those people and that information. If you’re lying, no one will retweet you. Even when people do fall for your falsehood initially, like the rumors which hit a critical mass in the study, the correction will be not far behind. Too many of these and you’ll lose credibility and followers. Followers and retweets are your cred on Twitter, so it is something to be careful with.

Speaking of bad information, new form of censorship has infested Twitter in Russia lately, hashtag spamming. Or as the article put it: “speech-canceling noise bots.” It’s not new, but usually used as spam advertising. If it gets thick enough, the hashtag becomes unusable. The #OccupyLA hashtag was spammed recently with psychic reading ads. I had to scroll through a dozen of them before finding a real tweet. The group switched to the #OLA hashtag, probably in response. It’s also shorter. Whether this was an economic spammer or an intentional tactical strike by OccupyLA’s opponents is unknown. Would be interesting to find out. Either way, it will be interesting to see how the twittersphere evolves to handle this threat.

That all said, I have two points to make. First is that for anyone wishing for an egalitarian, horizontally-organized society, these forms or networked democracy are the best real-world examples of what is possible. Forget quaint references to the Spanish anarchists of the ’30s or some rosy vision of pre-industrial tribal life or this or that Utopian commune. Here are your working examples, millions of people strong.

The other point is that as activists we have to keep this in mind when using social media in tactical situations. Disinformation can spread as fast as good information and in intense situations, people can imagine all kinds of things, leading to poor decisions in a rapid-fire environment.