Facebook is expanding fact-checking to photos and videos. The method will take advantage of machine learning to identify false content, which can be curbed from spreading. The move comes as the social media giant is struggling with the issue of spread of misinformation on its platform as well as user privacy. To tackle this, the company has is working with independent, third-party fact-checkers to review accuracy of content that is circulated on the platform. The social media giant said in a blog post that its fact-checking method for photos and videos is now available to all of its 27 partners in 17 countries around the world.
Facebook’s machine learning model takes into account several signals, including feedback from people to identify misleading photos and videos. The content is then sent to factcheckers for their review, who evaluate photos and videos using techniques like reverse image search. The metadata of photos and videos like when and where they were taken is also done by factcheckers. Other methods include using research from experts, academics or government agencies, as per the Facebook blog post.
“Based on several months of research and testing with a handful of partners since March, we know that misinformation in photos and videos usually falls into three categories: (1) Manipulated or Fabricated, (2) Out of Context, and (3) Text or Audio Claim. These are the kinds of false photos and videos that we see on Facebook and hope to further reduce with the expansion of photo and video fact-checking,” Tessa Lyons, Product Manager, Facebook said in a post.
Of course, false news spreads in different forms, which varies from country to country. For instance, in the US, misinformation is more common in articles, while it is misleading photos in Indonesia. Another method for fact-checking is to extract text from photos using optical character recognition (OCR). The text is then compared to headlines from fact-checkers’ articles to identify whether it is fake.