Facebook extends fact-checking tools to photos, videos: Here’s how it works
Facebook continues its efforts to curb fake news by rolling out fact-checking tools for photos and videos.tech Updated: Sep 14, 2018 13:36 IST
Facebook on Friday announced that it is extending fact-checking tools to photos and videos on its platform. These tools are already available for text and links on the social networking platform.
Facebook works with independent, third-party fact-checkers to review articles published on the platform. It has now expanded its fact-checking tool for photos and videos to its 27 partners across 17 countries.
Facebook has three categories for fake photos and videos. Content which is ‘Manipulated or Fabricated’ are used with intent to deceive users. ‘Out of context’ content are photos and videos which are authentic but misinterpreted.
The last category includes photos and video which have false text or audio.
How Facebook’s fact-checking tool works
Facebook uses machine learning to identify ‘potentially false content’ mostly through user feedback. Facebook then sends these fake photos and videos to its fact-checkers for reviewing. Tools used to fact-check photos and videos include reverse image searching which uses data to look for similar images on the web. Fact-checkers also analyse image metadata for details like location, date and time, device, file type, exposure, photo modification date and time.
Facebook says that its fact-checkers can “assess the truth or falsity of a photo or video by combining these skills with other journalistic practices, like using research from experts, academics or government agencies.”
A more advanced technology used by Facebook is its AI tool, Rosetta. This optical character recognition (OCR) tool extracts text from photos and videos and compares it to articles. Facebook says that using Rosetta AI it can process over a billion photos and videos in different languages per day.
First Published: Sep 14, 2018 13:36 IST