Facebook Messenger 'M' AI Assistant Now Snoops Your Chats To Offer Context-Based Suggestions

Facebook’s “M” AI assistant is getting even more powerful, and will leverage that power to provide “suggestions” as you use the increasingly popular Messenger app (over 1 billion active users per month). Facebook has been testing M’s interactions with Messenger on a “small percentage” of its users, and we are now seeing the fruits of that labor.

At its core, M is a machine learning AI assistant that scans your chat sessions with friends or family, and then provides relevant actions that you can either choose to interact with or completely ignore (you can also shut it off completely). Some of these actions are rather simple, such as suggesting a complementary sticker if you tell someone in chat “goodbye” or “I love you”.

facebook m

However, M can get into more involved situations, such as identifying when you are talking about paying for something (i.e. tickets that your friend bought, or repaying a friend for that gas money you owe them for the camping trip to the mountains last weekend). In this case, M would provide an option to send or request money.

Other actions that M can handle include location sharing, allowing you to setup a poll amongst friends if you’re indecisive on where to go for dinner (for example), and to even hail a car using ride-sharing services like Uber or Lyft.

“We are bringing the power of M’s AI technology to support and enhance the Messenger experience and make it more useful, personal and seamless,” said Facebook Product Managers Laurent Landowski and Kemal El Moujahid. “The more you use it, the more it can help. You always have the option to ignore or dismiss a suggestion if it’s not helpful, and if you don’t want M’s assistance you can easily mute it in M settings.”

M launches today for both iOS and Android users in the United States. Facebook says that it will “eventually” be introduced in additional countries as testing is wrapped up for other languages.


Via:  Facebook
Show comments blog comments powered by Disqus