Facebook is now training AI to review hateful memes to prevent them from creating chaos. This decision is taken as most of the staff are working remotely due to lockdown, which has forced only the machines to take up all the work. Facebook has already made a database of about 10,000 Memes which will help the AI to learn. As of now, Facebook has removed 9.6 Million posts that violate the community guidelines.
we are able to find more content and can now detect almost 90 percent of the content we remove before anyone reports it to us.
Guy Rosen, Facebook VP
The AI is backed by crowdsourcing with a motive to identify and execute malicious posts. According to Facebook a collection of about 10,000 memes has been made to train the AI. Facebook is releasing the database for researchers to work on. The initiative named as “Hateful memes challenge”, it will help the team to bring up an algorithm which can detect hate speeches. The winner will be awarded 10,000$.
These efforts will spur the broader AI research community to test new methods, compare their work, and benchmark their results in order to accelerate work on detecting multimodal hate speech
Facebook on a blogpost
