How Do You Know What Tweets are True? Exactly the Same Way You Source Everything Else
Sigh. NPR writes paragraphs on the difficulty of knowing which Tweets about Iran were true. Is it such a difficult question? You know what Tweets are true in much the same way you determine what non-Tweetable information is true.
Social media isn’t some separate sphere: it’s similar, in fact, to all other source building. If you don’t know who to trust, online or off, then you’re not doing your job. It’s what reporters do: they figure out who to trust before they need to. You should have built social media sources beforehand in exactly the same way you build offline sources.
Build Sources Beforehand
You know there’s an important Iranian election coming up. You know that the Iran government doesn’t support freedom of the press. So, it would make sense to build trustworthy sources online. Exactly in the same way you build sources offline. You find them before the news breaks. You dm them, email them, have lunch, talk about other issues to measure their level of judgment, find out who they are, build trust. Then, when they have news, you have a pretty good idea that it’s true.
Sure, there’s a chance that the Iranian government has murdered your source in order to use their Twitter account to spread propaganda, but it’s just the same as calling your source on the phone and realizing that their voice isn’t the same. If suddenly their grammar is different and they seem more strident that normal, you should worry.
If you don’t know what Twitter accounts to trust beforehand, you haven’t done your job, mainstream media.
You Verify the Information
If you haven’t built up trustworthy sources beforehand, then you need to focus on verifying the information. You look for other sources to confirm, exactly as you would offline. You look closely at the grammar, the spelling, the timing of the tweets, the details in the information. You dm the Tweeter to ask questions that make you believe in the veracity. You work from knowns: things that you’re certain are true. Does the Tweeter report accurately on these things?
You look for sources that aren’t ideologues: sources that are focused on truthtelling. Sources that have scientific and journalistic ethics based on telling you what is. You know when you find them, because such sources are clear about their biases, and still, will tell you information that doesn’t support their ideology.
Propaganda should be easy to spot on its own, without much verification. I think it’s unlikely that the Iranian government would say bad things about itself in order to spread a few untruths.
You hire people that are trained in verifying images. You hire people that are trained in analyzing textual patterns. You use technology.
And, you work the adversarial system the same way that it’s done in court: you ask the opposing parties to dispute what you’re hearing from the other. The side that does it best wins. You use the system to create truth a little more effectively next time.
In the End, You Just Say Who it’s Coming From
In the end, you can only verify so much. At a certain point you have to throw it out to the crowd. But tell the crowd in detail what you know, what you did to determine the veracity, and what else you’d need to know to determine the truth.
Because lots of the time, even in person, your sources seem dubious. Take this Bloomberg story for instance: they report that “a policeman” told Agence France-Presse that 1.5 million to 2 million people attended the Iran protests.
Sure, maybe he was a policeman. But why trust his report? You need to protect your source, certainly, but tell us how said policeman was measuring the size of the crowd. Tell us who he worked for; tell us who he voted for. Tell us at what size the protesters measured the crowd, and at what size the government did. Triangulate the truth.
In the end, isn’t the policeman’s statement not all that substantively different from a Tweet? Sure, you know he was there. But will that really give you better data? Maybe not. You also have to ask the policeman about the end and beginning points of the protest. Ask him how close the protesters are to each other. Ask him how fast the crowd is moving. Then do math to verify.
Look at it as a Fermi problem. Let us know that you thought about it. Can the streets possibly hold that many people? With that many people, how many square feet of street would the rally take up? Does this match the end and beginning points of the march?
In the end, all we want is to journalists to think about and disclose the reasons why or why not we should or shouldn’t trust the source. We don’t need their conclusion so much as the thought process.
Sources! Join Our Network!
We’re trying to build such a network. Of people we can follow beforehand and learn to trust: before the news hits. We’re always on the hunt for such sources, as described above. People that are objective, not ideologues: people that are invested, inherently, in the truth.
To build a journalistic institution ground-up for the Internet era, we like to get to know you beforehand. We have lunch, send emails, talk on the phone. We learn about what section of the world you know about. We learn to trust each other. We build expertise in knowing what you know. When the international event does come around, we’ll give you the latest undetectable technologies to dialogue with us. We’ll brainstorm about the best ways to get coverage. We’ll develop expertise in staying safe; we’ll do everything in our power to protect you, to support you, but we should be able to get a lot of information without putting you in danger: or at least any more danger than you are in doing your work there, on the ground.
The people on the ground are the people with information. The reporter simply needs to be able to access such people, protect them as sources, and build a relationship of trust. For instance, a low-level assistant to a city counsel member likely already knows more than a reporter will find out in weeks of digging. The people in Iran who are actually protesting know more than the reporters covering them. An aid worker in Gaza knows information far before reporters ever find out. The future of reporting is figuring out how to leverage technology to get trustworthy information from people such as this.
If you are such a source, do email us at network@ofrecord.com, introducing yourself.
Because a Western reporter plunked down in the middle of Iran isn’t the best way to get information. Nor is asking a policemen his estimate of the crowd without verifying it. Nor is live-blogging each random Twitter.
There’s no substitute, in the end, for the painstaking work of source-building among the skilled, both online and off.
