Google and Jigsaw have come together and launched a new technology called Perspective. This technology uses machine learning to find harmful comments and now it is available to the developers also, where they can use and build on top of this software. Jigsaw’s software is not a new one, it is there from past many years with different names.
Read more in Google Official Blog
Online trolling has become a common issue in the current social media friendly world.This software will be identifying the harmful or toxic comments that are made online. By mining a huge number of comment, a particular range has been fixed and presented to a panel of 10 people for their feedback on this. Demo version of this tool is available on the Perspective API website where anyone can use it by typing in comments and get to know how much abusive or toxic they are.
Certain quotes or phrases that have been working on mainstream in the recent months ranked high on the toxic scale as per the Perspective tool. For example, consider “Fake news!” is scaled to 47 percent that is very much similar to comments made by people are said to be “toxic/abusive.” “Bad hombre” is scaled to 55 percent that is labeled by people as “toxic.”
But there seems to be a slight refinement in the tools that not picking up. As it is a part of the learning process in machine learning.While testing this Perspective tool, developers tried some of the comments made on video was abusive was scaled to 3 percent similar to the toxic comments. In one of the video on YouTube, there was a comment -- “I have an eating disorder” which was scaled to 26 percent similar to other terrible remarks.