Even when one report a post or comment on Facebook for being too abusive and threatening, the troll content lie there for endless time. Unfortunately, most internet-tech companies struggle with this exact problem, should they even grapple with it.
While, the Wikipedia is researching ways to detect personal attacks against its volunteer editors other digital companies like New York Times, The Guardian and The Economist are either building an open source moderation tool or researching how best to moderate comment forums, host better and high-quality discussions & debate without an iota of abusive speech.
Now Google with its offshoot Jigsaw (formerly Google Ideas) has come up for rescue by launching a project called — Perspective — which aims to defeat online harassment using machine learning. Perspective was created by Jigsaw and Google’s Counter Abuse Technology team in a collaborative research project called Conversation-AI.
The team has now opened up this troll-fighting system to the world as an API that gives any developer access to the anti-harassment tools. Developers and publishers can use the scores evaluated by the API’s machine learning models to give real-time feedback to commenters or help moderators do their job, or allow readers to more easily find relevant information, as illustrated in two experiments demonstrated on its website.
Notably, this AI tech is still in its early stage and Google is intended to open source experiments, models, and research data to explore the strengths and weaknesses (e.g. potential unintended biases) of using machine learning as a tool for online discussion.
As AI is taking on human nuance it might not necessarily mean they’re always going to get it right. But as our internet discourse is quite toxic across the board, we’ll take any tools we can get toward making it a less hateful place.
[Top Image – Shutterstock]