Facebook has announced that it will, with its partners, begin fact-checking photos and videos in order to further stamp out misinformation. To date, most of the fact-checking has been oriented towards online articles, however, short videos or photos are also important avenues for misinformation to spread.
Discussing how the process will work, Facebook says it has built a machine learning model that uses engagement signals to identify potential misinformation. The content is then passed to fact-checkers who analyse content using various techniques including metadata analysis and reverse image searches. They will combine these skills with other “journalistic practices” including using research from experts, academics, or government agencies.
The input that Facebook receives from fact checkers will boost the capabilities of the machine learning software. In the announcement, Antonia Woodford, Product Manager at Facebook, said:
“As we get more ratings from fact-checkers on photos and videos, we will be able to improve the accuracy of our machine learning model. We are also leveraging other technologies to better recognize false or misleading content. For example, we use optical character recognition (OCR) to extract text from photos and compare that text to headlines from fact-checkers’ articles. We are also working on new ways to detect if a photo or video has been manipulated. These technologies will help us identify more potentially deceptive photos and videos to send to fact-checkers for manual review.”
The firm admitted that weeding out misinformation is a long term project, but in the short term it is investing in technology and partners to stay ahead of new types of misinformation that may crop up.