There has always been attention drawn to social media platforms, especially Facebook, and how it deals with misleading and harmful information. Perhaps this has increased recently, with more people spending time at home and on social media.
According to The Independent, Facebook is starting a competition, offering prizes to developers who come up with an AI algorithm that will be capable of identifying “hateful memes”. The company says memes are difficult to read by AI because computers don’t recognize contextual information the way humans do. So detecting an image that combined with a text and together makes up harmful materials, can be a challenge for AIs. If completed, the algorithm would be applied on all Facebook-owned platforms including WhatsApp and Instagram.
To give a hand to developers, Facebook created a series of memes, based on current ones being shared around the platform. The file consists of more than 10,000 examples of memes. The purpose is to identify the hateful memes so they can be deleted from the platform.
The competition has a prize pool of $100,000 and is hosted by DrivenData. Facebook also mentions that it’s making serious efforts to detect hate speech and harmful content that violates the company’s Community Guidelines. Mark Zuckerberg and Co hope this project will help Facebook keep people safe on its social media platforms.