As part of countering misinformation and fakes, Facebook is expanding its verification area to include photos and videos. Previously, the so-called fact-checking applied only to text publications.
Recall that Facebook is actively fighting false news and uses third-party verification for this.
Recommended: Jesica is an Instagram promotion app (from the creators of Instaplus.me). Install Jesica on your Android phone or computer and the app will like, follow, and unsubscribe as if you were doing it yourself:
- No blocks. The limits work as if you were doing it all manually.
- Safely. You do not need to enter a password for your account.
- High speed. You can put up to 1000 likes and 500 subscriptions per day.
Try 5 days for free >>
Advertising
On September 13, the company’s product manager, Antonia Woodford, announced that the social network is expanding the scope of verification for photos and videos to all 27 partners in 17 countries around the world.
How Facebook photo and video validation works
To identify content that could potentially be false, Facebook has developed a self-learning model. One of the signals that the photo is fake is the responses of ordinary users (comments, complaints).
If a published video or photo falls under any criteria of disinformation, the system sends them to its partners for verification – fact-checking…
Most fact checking partners have experience in evaluating photos and videos and are trained in visual verification techniques such as search original image and metadata analysis. For example, when and where the photo or video was taken. Fact checkers can assess the veracity of a photo or video by combining these skills with other journalistic practices, for example using research from experts and academics.
As accumulation of assessments photos and videos, the social network will improve the accuracy of the machine learning model. During content validation, other technologies such as optical character recognition (OCR) will be used to extract text from photos and compare that text with article titles. Technology is under development to help identify whether a photo or video has undergone any changes before publication.
Types of fake photos and videos
Based on months of research and testing with multiple partners, Facebook identifies 3 main categories of disinformation in photos and videos:
- Fabricated images.
- Headings taken from the context.
- Appeals and propaganda.
Earlier we wrote about 3 new Facebook split testing featuresand that Instagram is working on the launch of a new platform for online shopping IG Shopping…
We publish the most relevant materials in our Telegram channel: https://t.me/postiumru/. Go ahead and subscribe – no ads, just pure benefit!