Facebook Rosetta AI is a new technology which will help the social network find content that violates its policy guidelines…
It’s just not possible for Facebook to hire enough moderators to police all the content posted on the platform. That’s why the corporation is in the process of building new AI or artificial intelligence that will help out its human employees. The social property is creating AI called “Rosetta,” which uses machine learning to identify text placed over images and videos. More particularly the network is focusing this tool to target memes.
Facebook Rosetta AI to Identify Memes which Violate its Content Policies
The text transcription tech isn’t new but the challenge comes from the sheer volume. And, this is only complicated by differing versions of essentially the same thing. So, Facebook is unleashing Rosetta to extract text from 1 billion images and videos frames posted every day on Facebook and subsidiary Instagram.
At present, it isn’t entirely known the scope of what the company plans to do with this technology. It’s definitely useful for tools like photo search and screen reading. But, it could also be used to identify which memes are appropriate and relevant for users’ News Feeds. And, which ones violate its content policies.
Facebook explains it’s combining machine learning with text extraction to “automatically identify content that violates our hate-speech policy.” This is deployed across several different languages. In light of recent scandalous events, this could prove very useful for the social network.