Facebook’s new DeepText AI will be capable of not only analysing all input across the social network, but will assign contextual adverts and audiences.
Facebook has been exceptionally public in its research into and eventual deployment of bots within Messenger, and now the company’s DeepText AI is set to examine every shred of writing across the length and breadth of the social network.
DeepText is set to help Facebook analyse several thousand of the 400,000 new stories and 125.00 comments that arrive per minute on public posts. Read: Facebook wants AI to help them understand your content
The AI will first be noticed on Messenger; for instance, if DeepText identifies a user may be in need of transport through analysing a message such as “let’s take a ride there”, the intelligence will offer up a quick link to either Uber or Lyft – both of which have integrated into the platform.
Facebook has explained the intention of the artificial intelligence, noting that for the company to “get closer to how humans understand text, we (Facebook) need to teach the computer to understand things like slang and word-sense disambiguation. As an example, if someone says, “œI like blackberry,“ does that mean the fruit or the device?“
Facebook has further highlighting its AI’s scans of personal messages, which has previously made the company the target of a class-action lawsuit. However, the AI will also be used to help the social network identify harassment and abuse, citing that its AI systems already report more offensive photos than actual human beings do.
Using DeepText, Facebook hopes to improve the way in which it shows content from certain authors and topics to different users; giving the social network a massive boost over services like Twitter which are presently changing to an algorithmic timeline. Read: Mark Zuckerberg aiming to build his own AI assistant
What are your thoughts on Facebook’s new AI system? Be sure to let us know your thoughts in the comments below!
Follow Bryan Smith on Twitter: @bryansmithSA