Here's Facebook's vision for the future of AI

Facebook is investing heavily in what many in the tech industry believe to be the next frontier of innovation, artificial intelligence. And it's already using AI to dissect what's in photos and videos.

The largest social network on earth has a division of AI experts it calls FAIR. There's also a separate team called Applied Machine Learning, which focuses on "giving people communication superpowers through AI."

There are a few key areas Facebook is using AI in right now:

• Filtering the News Feed: There are potentially thousands of stories that could be shown to you in the News Feed at a given moment, and the company uses AI to learn from what you like and show you stories that are more relevant to you.
• Language translation: Facebook says that 50% of its users don't speak English, and it's using AI to translate text on a massive scale.

Facebook

• Photo search and tools for the visually impaired: Facebook says that AI is currently making the most progress in computer vision, which uses artificial intelligence learnings to analyze what's in an image. The idea is that you could search for "snow" and see photos with snow in them. (Google Photos does this already very well). Facebook also uses AI to read aloud what's in a photo for the visually impaired. 
• Analyzing what's in videos on Facebook: Given how increasingly prevalent video is on Facebook, the social network is trying starting to analyze the contents of videos in real time and assign relevant content tags based on what's in them. 
• Training its digital assistant, M: Facebook has been testing its own chat-based digital assistant called M inside the Messenger app. M uses a mix of AI and human trainers to help you do everything from buy movie tickets to research restaurants for dinner. M won't be ready for prime time for awhile, unfortunately. Facebook is currently testing it with a few thousand beta users in California.

Facebook clearly believes that AI is important to the company's future. Its employees are running 50x more AI experiments per day compared to last year.