Over the past 15 months, there’s been a huge movement to automate the removal of extremist based behavior that’s been plaguing our favorite social networks.
The New York Times, TechCrunch, NPR, The Information and many other notable publications have chimed in as this being one of the current major problems in technology to solve over the coming years. Extremist groups, trollers, homophobic, racists and people who are simply filled with hatred, have taken to these networks to lash out at civilized, professional and otherwise friendly consumers. Twitter and Facebook the two largest and thus most vulnerable networks for communication have recognized this as a major concern.
At HappyFunCorp we viewed this as both something we could help solve-to better humanity on the internet and also, a complex technology challenge. How can we build a framework that enables the automatic removal of extremist behavior while not obstructing the day to day users communication?
Like most complex problems, at a high level, it appeared to have a fairly simple solution. However, after coming out of the technology product hole, we can say with certainty, initial assumptions were altered.
KNOWMORE uses a blend of machine learning, string analysis, (in this case tweets) and sentiment analysis from a master ‘block-list’ to provide a harassment free experience. Over time, the master block list gets shared deeper across the graph as more people use KNOWMORE and thus gets smarter and smarter. For the user, we’ve built a tolerance threshold defined by our dial, where the user can open up his or her stream to let everything in or, dial in back to avoid any hatred filled communication.
After a number of product revision, user testing and more product revisions, we’re pleased to invite twitter users to sign up for our Beta. We are keeping a very close eye on the algorithm so we will only be letting in ‘small’ groups at time. As the algorithm gets smarter and the machine learning refined, we’ll begin rolling out support for Facebook and other large networks.
We believe our solution closely coupled with users and the help of teams inside respective networks, can and will substantially move the needle in regards to extremists groups harassing and taking advantage of people online.
This is a solvable problem.
The core team is Jon Evans, Will Schenk, Aaron Brocken and Milos Roganovic