Google Launches New Tool For Fighting Trolls, Shares It With Developers

Trolls are practically an accepted part of the internet now, with many in various comment sections saying “Welcome to the internet!” to anyone who expresses astonishment at offensive statements on anything. It seems that Google doesn’t like the status quo, however, and is determined to stamp out abusive behavior on its platform. With a new artificial intelligence targeting specific words that it is sharing with other developers, it might just be able to pull this off.

As Forbes points out, this approach is basically fighting the words that are used by the trolls instead of engaging them in a war of ideas, which is always a mistake. Called Perspective, it’s an algorithm that’s propelled by advanced deep-learning technology. It was fed with over 17 million comments and phrases, which was compiled by The New York Times.

These sentences were then rated based on how toxic they are, with numbers going from 0 to 100 percent. Depending on what words were used and how they were used, the algorithm would identify the characteristic vocabulary of a habitual troll.

What really sets this method apart is the fact that it goes after comments that are meant to cause emotional abuse through offensive words. Trolls could still potentially ruffle feathers, but now, they would have to actually articulate their thoughts in a clean and clear manner. Anyone who has ever come across a typical troll will know that this is practically impossible for them to do.

Soon, Perspective will be shared with other developers so that they too can use it on their own platform, Computerworld reports. The most likely companies to make use of this tool are those involved in content sharing that also has a comment section since these are often where trolls thrive. Of course, the developers would still need to be screened to get access to the tool.