Facebook might go for artificial intelligence to curb fake news and videos depicting violence as they are reportedly figuring out how to introduce AI responsibly.
Facebook Inc. has said that it is contemplating using artificial intelligence to deal with one of its biggest issues; that of fake news infesting its News Feed. However, a successful implementation of the idea is still dependent on the policies that the social site comes up with to make the best of AI technology, the WSJ reported.
Apart from fake news, videos depicting violence or other objectionable content are also on the radar of social networking site. However, as Yann LeCun, Facebook’s director of AI research says, the company is yet to frame the right policies that would lead to successful and responsible implementation of artificial intelligence.
Even the unexpected victory of Donald Trump in the recently concluded the U.S. presidential election is being blamed by many op-eds on fake news. This again should serve as a measure of how much of an impact that fake news might have on the functioning of Facebook, or for that matter, the society at large given the immense popularity that the site enjoys.
No wonder, Facebook’s inability to effectively weed out fake news has also attracted the ire of numerous publications that has pinned the blame on an algorithm that the site uses to lists news stories.
The same also applies to the video content posted on Facebook that might end up depicting violence. However, the site right now depends on other users’ flagging such content to limit the spread of such videos. That might change once AI comes into the picture though as Joaquin Candela, Facebook’s director of applied machine learning said, it comes with its own set of challenges.
For instance, getting AI to flag objectionable video will require them to have extremely fast computer vision. Then the AI also needs to have the right filtering algorithm to prevent it from pulling down the right ones.
The same also applies when it is about fake news as the AI will need to know how to differentiate genuine news stories from what has comes to be known as the click bait articles.
The above issues are compounded further given the sheer size of Facebook and its operations. It is no less than 1.79 billion people who access Facebook at last once a month with more than 1.57 billion accessing the site from their mobile devices. Combine that with the number of countries and languages the social site is available in and you will know how big a scope the AI algorithm needs to be.