Artificial intelligence is leading the way everywhere. After Google’s AlphaGo AI set a world record by beating the World Go champion, this time it is Facebook’s turn to do something that involved two words — Out of the Box and AI. The company has announced the development of an artificial intelligence system capable of analyzing what people are writing about in their news feed posts.
Christened DeepText, the AI possesses a near human accuracy along with proficiency in over 20 languages. The company hopes that the AI will be able to pinpoint posts and updates that spew poison, thus enabling Facebook to remove them before they are able to cause any damage.
If it can automatically pick out hate speech from status updates that merely contain controversial terms, it could prevent a human from ever suffering the harassment and having to report it manually.
The development of the AI was not as simple as it sounds. I mean there is a whole world of algorithms involved in teaching the AI to differentiate between “This is my phone, This is my phone and This is my phone — Child’s play for humans, but very difficult for a computer program.
However, DeepText comes with capabilities that enable it to do far more — including understanding when a person is indulging in hate speech. Facebook already has an automatic image reporting system that reports more images every year than humans and DeepText will just add another tier to its security.
But reporting is not just what DeepText limited to. The AI can also offers suggestions based upon your situations. So if you are talking about needing a ride, it will help you out with Uber or Lyft. Or if you are looking to sell or buy something, it will automatically come up with information such as the object being sold and its price quotes across the web.
So yup, very handy! The service is being tested with Facebook’s messenger service as we speak. However, the company is planning a deeper integration of the DeepText AI with its platform.