
Google has just announced the start of the activity of its own troll detector.
Google understood the problems the modern society is fighting with and took a proactive step into combating them. As some of these issues regard attacks on the right to online privacy and security, the tech giant felt it was part of its responsibility to turn the digital world into a prosperous place for all kinds of communities. This is why Jigsaw unit that is part of Google announced the creation of a new troll detector that will also be open source.
On Thursday, Jigsaw unit announced a collaboration with other members of the Counter Abuse Technology Team at Google. The result of their work is the Perspective project that uses a software built on machine learning technology. This piece is going to become an online troll detector that can identify abuse and harassment in the online medium. The Perspective project is not new. Google created it in 2010 and used it to discover ISIS recruits. However, the API is modified so it can reach its new purpose and allow outside developers build their own projects on it.
The software uses a toxicity scale to detect possible harmful comments. Its machine learning technology went through millions of comments throughout the digital world. Afterward, the AI presented the comments to a board of ten people who ranked them. Online users can employ a demo version available on the Perspective website where they can insert any comment they want to scan for its potential harmful nature.
The troll detector is already guarding major comment sections against any harmful message. Publishers such as The Guardian, the Economist, and The New York Times have already implemented this software. By doing so, they want their comment sections to become a place for clever debates that do not rely on insults to make a point.
CJ Adams, the Conversation AI product manager, stated that the future of this technology would create a safe online medium where conversations are meaningful and convey a personal, honest, and clever message. He also sees the troll detector as a new way to reactivate comment sections among publications that had no other choice when it came to dealing with users’ negative comments. The main threat Adams saw in the modern world is that toxic comments push people to interact with only those persons that share their views. This is a bad social phenomenon, as users are creating their own bubble where no different perspective can enter and show them where they are right and where they are wrong.
Image source: 1