Somebody Is Fixing Twitter's Misinformation Problem, but It's Not Twitter

The Atlantic

View gallery

.
retwact-04.png

Even worse than the pornbots and trolls, one of Twitter's enduring drawbacks is that there's no reliable way to issue corrections with the service. The flaw drew a truckload of attention in the wake of the Boston bombings as self-appointed detectives -- fueled by Reddit -- made all kinds of invalid allegations on social media. But the tragedy involving the Tsarnaev brothers is simply the most extreme case of what's a very ordinary problem. Virtually all reporting has become an iterative process, subject to revisions in real time as new knowledge becomes available.

Some journalists (including me) have called for Twitter to build a corrections function involving the same code it uses for retweets and favorites. So far, Twitter has given no indication that it plans to do so, which has led one independent developer to build his own solution. And until something more polished comes along, it's not a bad option.

The tool is called Retweet Retract -- or Retwact, for short. Built as a side project by Stonly Baptiste, a software developer at the Pennsylvania-based company independenceIT, Retwact tries to contain the damage (and shame) that comes along with spreading information that later turns out to be untrue.

Here's how it works. Once you've given it permission to look at your account, the app culls your five most recent tweets that have garnered retweets.

View gallery

.
retwact-02.png

Click on one of them, and a little green radio button fills in. Down below, you can tap out a hand-written apology or send the correction as-is. When you publish the notice, it goes from your account to all your followers. Meanwhile, Retwact's own Twitter account (@Retwact) reaches out with an @-mention to the first 100 people who retweeted your (incorrect) tweet.

The link above takes you to a side-by-side comparison showing both the older, incorrect tweet as well as the newer, corrected one.

Baptiste told me he hacked the tool together in a "sleepless" 48-hour session, using resources he'd never worked with before such as OAuth and the Twitter API. That was after he'd posted the idea to the social news aggregator Hacker News, where it got promoted to the front page within an hour and received upwards of 1,500 requests for Retwact to be built. (Twitter, are you listening?)

Retwact isn't perfect. If you tweet often, you can only use the service to correct something you tweeted minutes rather than hours ago. And Retwact can't detect manual retweets of misinformation -- just the ones produced when you click Twitter's own retweet button. But Baptiste said he plans to keep working. He's considering adding a feature that will delete the original, misinformed tweet altogether while still preserving the text on the correction page. That means there'll still be a record of the mistake, but other users will no longer be able to spread it around.

Baptiste promises to improve the app until "A) People stop using it, or B) Twitter (or someone else) offers me a lot of money for it."





More From The Atlantic
View Comments (0)