Read on Twitter
Thread: How social media companies are fooling us all regarding their failures to fact-check issues on social media.
I want to use this simple example to explain what social media companies can do and are choosing not to.
I want to use this simple example to explain what social media companies can do and are choosing not to.
An account with 10 followers tweeted about a missing uncle in the middle of the strife going on in Minneapolis, US, and was retweeted 14.5k times. This is a common strategy to spread misinformation - Anonymous account + image or video + emotional story during a tense situation.
The image is an old image and was featured on the website of Star Tribune in November 2017.
https://www.startribune.com/the-world-of-the-pictureman/457343063/
How could a social media company like Twitter have prevented this image from being used as part of a misinformation drive?
https://www.startribune.com/the-world-of-the-pictureman/457343063/
How could a social media company like Twitter have prevented this image from being used as part of a misinformation drive?
To explain that, let me explain how easy it is to do this search. Here's what I did on my laptop. My primary browser is Firefox, and I use a Firefox addon called 'Image Search Options'.
https://addons.mozilla.org/en-US/firefox/addon/image-search-options/
See the video.
https://addons.mozilla.org/en-US/firefox/addon/image-search-options/
See the video.
I used a popular technique called 'reverse image search' that is provided by many search engines. The most useful implementations are by Google and Yandex. For those who want to learn a few techniques of how to spot misinformation, try this free course. https://www.poynter.org/shop/fact-checking/handson-factchecking/
Question is, do Googles, Twitters, and Facebooks of the world need fact-checkers to tell them or tell people at large that this is an old image? They don't. They can automate this process. Whatever I did in the video, can be done through a program without human intervention.
Google has a product called 'Vision API'. This product has been used by many people to automate reverse image search for various projects. Read this interesting research done using Google's Vision API. https://www.forbes.com/sites/kalevleetaru/2019/06/02/using-google-vision-apis-reverse-image-search-to-find-television-news-images-on-the-web/#71d6c90821b0
Google+Twitter+Facebook etc have very deep pockets, and it wouldn't be difficult for them to create a product which works across platforms which alerts users when old images are used to represent present events. And with some work, this can be extended to videos as well.
I'm not saying that this is going to bring an end to misinformation. And the automation cannot be done for all images/videos. But it is certainly possible to add contextual information to many images/videos that are being misused on social media.
If and when these social media companies decide to do this, it will neutralise a very common strategy used by those who create disinformation. Half of the @AltNews articles have the title - "Old image/video passed off as".
Why aren't they doing it? That's a subject for another day, but it all boils down to business. Meanwhile, if you want to not be constantly triggered by misinformation, lets start holding the social media companies more accountable.
