Can an Ex-Googler’s New Startup Get Rid of the Trolls?

A former Google exec's new social network for women hopes to keep out the trolls with A.I. and human editors.



In a recent survey, Pew Research found that 40 percent of U.S. adults have experienced online harassment in some form. While men were more likely to be called offensive names, women were more likely to be stalked and sexually harassed online. A new social network is hoping to to provide a platform for troll-free discussions.

VProud is the brainchild of Karen Cahn, a former Google, YouTube and AOL exec. She was formerly the general manager of AOL’s original content, and headed up sales, partnerships and branded entertainment for YouTube. The site bills itself as a “video-based conversation experience built for women, by women.”

VProud has attracted 250,000 users during its first month of beta. The platform allows women to start or join conversations around certain issues. The site looks more like a video-centric xoJane — a reflection of its staff. At launch, there was a managing editor, parenting editor, associate editor and a couple of contributing editors. Its staff page looks more like that of a media startup, making the entire operation a weird publication-social-network hybrid.

While publications are social with their comments sections and social-media integration, and social networks are news sources filled with shared links and user-generated content, VProud takes this integration to a new level. Its homepage has news, health and parenting verticals, and the social aspects of the site are centered around its verticals. Users who start their own conversations must include a headline and video, but the conversation is edited by a “team of experts” who edit conversations to “fit to the site’s aesthetic,” according to a how-to video.

From a release announcing its launch:

With online attacks against women increasing in recent months, VProud offers women a safe digital sanctuary to communicate with other women, free from verbal hostility and abuse. Thousands of topics are discussed in real time, including parenting, business & careers, education, healthcare, modern feminism, sex how-to, LGTBQ issues, addiction, mental health, fertility, pregnancy, body image, relationships, and more. Nothing is off the table, except for hateful and disrespectful speech, which is filtered out through a proprietary mix of technology and community involvement.


So how does the site keep the trolls away? To begin with, all users must sign up with a real-world identity, which ties her account to a Facebook, Twitter or Google+ profile. But users can choose to comment anonymously, by toggling between their real identities and an anonymous handle.

The site has three ways for keeping out the trolls. New York reports:

The first is an artificially intelligent database that filters posts that include name-calling and hate speech, understanding context and getting smarter over time. The second is community moderation. All posts have a one-click troll icon for flagging unproductive commentary. If a post gets a certain amount of “trolls,” it automatically gets taken down. Then it goes before a jury of human editors and community managers because “sometimes an unpopular opinion will get ‘trolled,’ but, frankly, it’s just an unpopular opinion,” Cahn said. If so, the comment will be granted immunity and — No. 3 — enter the A.I. filter, making it smarter still about context.

The three-step system seems like it could be very effective in keeping out abusive comments, and the toggle feature between public and private is useful for encouraging discussion around more sensitive topics. However, new conversation topics start with a video — a not-so-private mode of communication.

Another feature of the conversations are the Yes/No/Not Sure options when commenting. The labels could result in an oversimplification of the discussion. After all, not all points of view neatly into the aforementioned categories. Kat Stoeffel writes in New York that this “emphasizes argument over information and entertainment.”

It can, however, give interesting insights on what people think about a controversial topic. This week, a video went viral showing a woman getting catcalled on the streets of New York. In discussing this video, most women on VProud had negative views street harassment. But the No/Disagree tags highlighted dissenting opinions, like those who admitted actually liking catcalls:

Whether or not the new site will be able to successfully root out trolls remains to be seen. The process has potential, especially if the A.I. filter is effective. Another hurdle is building a significant user base: Will women really want to upload videos of themselves talking about personal issues?

Recommended articles