For anyone lucky enough to be unaware of the term, trolling
is the indelicate art of offending Internet users by posting inflammatory content (usually in comment sections, chat rooms or web boards) in order to create an emotional response or generally disrupt the conversation. The prevalence of trolling has led many Internet publishers to close their comment sections altogether
, and remains a constant headache for moderators for all sorts of Internet venues. With Perspective
, Alphabet and Google want to put a damper on that kind of toxic behaviour.
Alphabet’s technology incubator, Jigsaw,
handled the creation of the Perspective API. Jigsaw’s other projects include DDoS attack repellant, a web proxy extension for people living in repressed countries and AI software intended to spot the language of abuse and harrassment.
Jigsaw pushed hundreds of thousands of comments labelled as ‘toxic’ into Perspective. Using this huge base of examples, Perspective flags comments that would drive people away from online conversations.
The New York Times
has already tried using Perspective to mitigate their comment section woes — hopefully, this will allow them to open up more than just two percent of their articles to public commentary. Wikipedia, The Economist and the Guardian have also tried out the tool.
Google is maintaining a hands-off approach to the language standards of anyone who wishes to use the tool — each publisher will have their own definition of what ‘unacceptable’ language might be. It will be interesting to see how this kind of auto-moderation of comments online could redefine what people view as toxic.
Hopefully this will help relieve the psychological weight
human moderators currently carry while policing Internet discussion and flagging violent or disturbing content — as well as reduce company costs for this kind of 24 hour human supervision.