Jul 11, 2016
Verify
This section covers:
- Fake accounts
- Verifying images
- Verifying video
- More resources
Update: The US election in November 2016 and the debate around fake news has thrown a huge spotlight on this important area. For more reading and the latest updates, check First Draft News.
One of the most important things to remember when working with social media is that nothing is verified. Rumours spread like wildfire, and images are very often taken out of context (knowingly or unknowingly), or manipulated. Be suspicious!
There are some great guides and resources on verification out there, and wherever possible I link to those. There is a also a list of more resources at the end of this section.
N.B. This symbol ✳️ means the tool is free
Fake accounts
Few people realise how widespread fake followers and so-called “bots” (“robot” accounts that are programmed to behave a certain way) are, and how easy and cheap they are to buy.
It’s a curious kind of grey area — against the guidelines of social media companies, but not actually against the law to buy or sell.
As Jeff Elder demonstrates in this excellent Wall Street Journal article, a whole industry has arisen to sustain this, which is tapped into by people from all walks or life — from celebrities to struggling businesses.
Often people buy fake followers just to make themselves appear more successful and popular, but there are also much more pernicious uses.
The Russian government, for example, has been shown to use both an “army of trolls” and an extraordinarily high number “bot” accounts in an attempt to skew the online narrative.
How to check if a Twitter account is real
The first thing to do is to look yourself carefully at the account.
When was it set up? Who does it follow? Who is it followed by? Does it have a profile picture? How often does it tweet? What does it tweet? Is it verified by Twitter with a blue tick?
This will give you a good idea if it’s genuine or not.
Some free tools
There are also some tools that can give you a further indication, but this is not an exact science, and none are full-proof:
✳️ Klout This rates accounts out of 100 for trustworthiness and influence.
✳️ Bot or Not This analyzes your own Twitter followers and gives an estimate of how many are genuine or fake
✳️ Status Media Fake Follower check This also analyzes your own Twitter followers and estimates how many are fake. You can also run the same check on up to two other accounts for free
✳️ Social Bakers also has a free fake follower check
✳️ Foller.me allows you to check some basics on an account very quickly — when it was set it up, which accounts it mentions most, what words it uses etc
An important thing to bear in mind is that someone who has a high percentage of fake followers may not be responsible for that themselves. They may have been bought deliberately to discredit them (sounds far-fetched, but can happen) or it could actually be a sign that the account is regarded as particularly trustworthy — fake accounts are often set to follow high-status genuine ones as a way of throwing people off the scent.
Verifying images
There are three indispensable tools for checking the origin of images — all are free.
- ✳️ Google Reverse Image Search
- ✳️ Tin Eye
3. ✳️ Jeffrey’s EXIF Data viewer
Google Reverse Image and Tin Eye do the same job — searching the internet to see if an image has been used before.
If you amend the settings to show the largest version of the image available, that will usually be the original.
To do a Google Reverse Image search, right-click on an image and select “search google for this image” or drop the image into the Google Images search bar. The Chrome extension is worth downloading.
To search on TinEye, it’s the same process. Again, selecting the largest version of the image will most likely lead you to the original. The Chrome extension is here.
Jeffrey’s EXIF Data viewer reveals various bits of metadata associated with an image, including what camera it was taken, the time and date, and — sometimes — the location.
This does not work for all images. Twitter, for example, strips the EXIF data when you upload a photo.
Other similar tools are Fotoforensics.com, Findexif.com and Photo Mechanic
All these tools are very useful, but as my BBC colleague Trushar Barot writes in his excellent chapter in the Verification Handbook, one of the best ways of checking the validity of an image is to talk to uploader, and ask a few basic questions:
- Who are they?
- Where are they?
- When did they get there?
- What can they see (and what does their photo show)?
- Why are they there?
Trushar continues: “One important aspect to note here: If the image is from a dangerous location, always check that the person you are talking to is safe to speak to you. Also be aware of any issues about identifying the source through any details you broadcast about him or his images.
“From our experience at the BBC, people who were really there will give visual answers, often describing the details in the present tense. (“I’m in the middle of X Street; I can see and hear Y.”) The more vague the answer, the more caution you should exercise about what the source is telling you.
“Another useful technique is to ask the person to send any additional images shot at the same time. It’s rare that someone takes only one picture in a newsworthy situation. Having more than one image helps you learn more about how the events in question unfolded.”
Another quick check of an image that’s been tweeted is to see what they used to tweet it. Find the tweet in Tweetdeck, then click on the bottom where it says “details”. That tells you what device/client they used. If they claim they are at the scene of a breaking news event but are tweeting from a desktop that might raise a flag.
A few other photo resources:
✳️ izitru This site analyzes images to test for their veracity, and then gives the image a “trust rating”
The Twitter account ✳️ @PicPedant is constantly on alert for fake images and images taken out of context
And coming soon is the ✳️ Verified Pixel Project which aims to simplify the process of image verification
Verifying video
They are now undisputedly the best resource for journalists working on verification, and publish regular updates and best practice guides in English, Arabic, Spanish, French and German.
In brief, here are a few key points when verifying video — again all are free ✳️
- Get in touch with the uploader
- Check the upload date. In YouTube, just under the search bar is a button marked “filters”. This allows you to order the results by upload date. The YouTube Data Viewer from Amnesty’s Citizen Evidence Lab can give you the exact time of upload
- Check the thumbnail. Do a reverse image search in Tin Eye (as above)
- Read the description carefully
- Look at the account carefully (What kind of videos does it post? How often? When did it open? Are there any contact details? Are there other social media accounts using the same username?)
- See if the video has been shared elsewhere. Take the video’s unique code (the bit after the v= in the url) and paste that into the search bar in Twitter and Facebook
- Crosscheck on Google Maps, and zoom in on Street View. Wikimapia.org is another option.
- Look for visual clues like the style of road signs, traffic lights, shadows which give an idea of the time of day, and listen out for clues such as people’s accents
- Check the weather when the video is said to have been filmed and see if that matches. You can do that at WolframAlpha
Update: Here are a couple of new Chrome Extensions that are very helpful for video verification:
Info on both of these is available at First Draft News.
There are some excellent guides on this, which I link to below. The chapter on video verification by Malachy Brown in the Verification Handbook is particularly good.
More resources
Reading:
✳️ First Draft News is now the go-to place on anything to do with verification or social media newsgathering. They publish news, best-practice and case studies in this area in English, Arabic, Spanish, French and German.
✳️ The Verification Handbook (also available in French, Spanish, Arabic, Portuguese and Ukrainian)
✳️ The Verification handbook for Investigative Reporting
✳️ Bellingcat is a hub for investigative journalism by citizen journalists run by Eliot Higgins, described as a “modern Sherlock Holmes”. It includes lots of useful how to guides and examples
✳️ Amnesty International’s Citizen Evidence Hub is a useful resource with lots of tutorials on various aspects of verification, and is home to the YouTube Data Viewer
✳️ Social Newsgathering: A collection of Storyful Blog posts edited by Claire Wardle. Came out in 2013, but is still full of lots of useful info
There is an extensive list of tools and resources for verification here
Forums and debunking:
✳️ Storyful’s Open Newsroom is a Google+ community where journalists and researchers are invited to help verify content, and share information
✳️ Checkdesk is a collaborative verification platform
✳️ Emergent a real-time rumour tracker founded by Craig Silverman
✳️ Snopes aims to debunk misinformation on the internet
Online training:
Getting It Right: Accuracy and Verification in the Digital Age by Craig Silverman for Poynter News University (costs $30, three hours)
Fact-checking and Debunking: Covering Rumors, Hoaxes and Gossip by Craig Silverman for Poynter News University (costs $30, one hour)
Online and Social Media Corrections: When, Why and How by Craig Silverman for Poynter News University (costs $30, one hour)
✳️ More from the Social Media Reporter ✳️
- How to organise your feeds
- Locating video, images and sources from a specific location, and the ethics of using eyewitness material
- How to use social media to track people down and for research
- How to find out what’s trending and dig to the bottom of trends
- More resources
This guide is a starting point. I’d love others to contribute to expand and improve it. Please feel free to leave a note here, or get in touch at cordeliaheb@gmail.com or @cordeliaheb
