While the researchers at ASI wouldn’t consider any technical specifics of the algorithm, it shows to work like other forms of A.I. recognition software. The algorithm can examine any video and determine the probability that the video is a bit of extreme propaganda. According to the News, the algorithm was practiced on thousands of times of terrorist selecting videos, and it uses characters from these videos to allow probability scores.
If a video is considered as very high probability, it is held for review by a human content moderator. Because the videos aren’t automatically considered down, any false positive should be caught before the video is wrongfully dismissed. ASI said that the algorithm could identify up to 94% of Islamic State uploads.
The algorithm, which was partially supported by the British government, was essentially created for the benefit of smaller video programs that don’t have the resources to maintain a large content moderation staff. Because these sites can’t completely examine every video uploaded, it’s easier for terrorist groups to plant promotion on them. As researcher Marc Warner of ASI told the News, “There are over 1,000 various videos on over 400 various platforms.”
It’s not clear when the new algorithm will be extended on these sites. When it finally shifts available, groups like the Islamic State will likely try to improve the way they craft their videos in order to avoid exposure.
Take your time to comment on this article.