Silicon Valley begins putting cyberbullies in the crosshairs

With the rate of digital bullying increasing, tech firms escalate efforts to build automated tools that can detect and flag online harassment. 

Melissa Hourigan knows first hand about the kind of online vitriol that kids encounter. She works in digital media and volunteers for a suicide prevention organization in Denver. But even that didn't prepare her to respond when online bullies began targeting her 15-year-old daughter.

In January, a fellow student began tormenting Ms. Hourigan's daughter on Snapchat and Instagram. But blocking the bully didn't stop the problem. On Valentine's Day, classmates told the teenager that insulting messages popped up on ThisCrush.com, a site that many young people link to their Instagram pages to send secret love notes.  

Ever since, Ms. Hourigan's daughter has become unusually distant. A skilled drummer and bass player, she didn't audition for a new band. She's ditched makeup and dresses for baggy clothes. The honors student's grades have even suffered. 

"It's been a disaster," Hourigan says. "She kind of became withdrawn, but the way that we've seen it materialize is the inability to focus at all. It's so drastic, she went from an A and B student to having four F's. We've had to get a tutor and counseling."

For victims of digital abuse, there are often few options to stop online bullying. Hourigan, a digital communications consultant, printed out the offensive messages and contacted the school's principal, the girl's parents, and local police. The tormentor eventually got suspended from school. 

Yet after the ordeal stopped, ThisCrush administrators appeared to rebuff Hourigan in a brusque exchange on Twitter, even though the comments appeared to violate the site's terms and conditions.

"You can call the police, which we did, and even they can just document it," Hourigan says. "But until something really happens, there isn't a lot to protect kids from things that happen on these sites, especially those that allow anonymous posting." 

ThisCrush did not respond to multiple requests for comment.

Concerned parents may not be able to convince developers of fringe apps that quickly go in and out of favor to help thwart bullying. But most established social media companies are beginning to take online harassment seriously.

In fact, many of Silicon Valley's biggest tech companies appear to be making a concerted effort to crack down on harassment on their platforms and news sites, a push that's motivated by the increasing prevalence of harassment. According to a November survey from the New York-based research institute Data and Society, nearly 47 percent of US web users have experienced some form or online harassment. 

Last month, Jigsaw, a technology incubator first launched by Google, announced the release of Perspective API, a beta version of a platform that could help news sites automatically ferret out vulgar and disrespectful online comments.

To create the new filtering software, Jigsaw hired a team of "human evaluators" to rate comments from The New York Times, Wikipedia, and other sources for "toxicity" – one measure of abuse. The Times, the Guardian, and the Economist, plan to test out the algorithm, which offers an easy-to-use interface for moderators. Type in any keyword or phrase, and Perspective will show you how many similar comments were marked abusive by internal judges. 

“Because of harassment, many people give up on sharing their thoughts online or end up only talking to people who already agree with them,” said CJ Adams, a product manager at Jigsaw. The goal of Perspective, he added, is "to help people stay online without having to read every bit of abuse hurled their way.”

And it's not just Google that's deploying next-generation tools to deal with the issue.

Twitter is also rolling out new tools to combat the rise of online bullying. In updates announced earlier this month, the company now lets users block specific accounts, keywords, and usernames from their timelines. Facebook also offers a "Bullying Prevention Hub" that provides tips to users hoping to thwart harassment. 

The issue is also getting more local and national attention. All 50 states have laws or policies addressing bullying. The subject made national headlines last year when an 18-year-old Texas girl shot herself after being the target of online attacks. And it even became a focus for First Lady Melania Trump as she supported her husband on the campaign trail, calling mocking, bullying, and attacks "absolutely unacceptable when it is done by someone with no name hiding on the internet."

Yet even though social media platforms such as Google, Facebook, and Twitter have struggled to root out online abuse on their platform, some digital experts remain skeptical about using artificial intelligence to solve the problem. 

In a post on Medium on Feb. 23, machine learning designer Caroline Sinders criticized Jigsaw's new platform, showing that the algorithm proved more likely to mark a comment containing the word "Arabs" as abusive than one that said "I love Fuhrer," a term of endearment used for Hitler.

"This is a data set of agitation and anger, and perhaps, not really toxicity," Ms. Sinders wrote.

Google has said that it's still tweaking the Perspective algorithm. 

So far, Hourigan, the Denver mother, likes what she sees coming from Silicon Valley to combat bullying. "I have been addicted to finding out what's possible," she says. "It’s a conversation that people [must] have."

But, she cautions that in a world where taunting can spread from social media platforms such as Facebook, Twitter, Instagram, and Snapchat to smaller apps and sites, the problem is becoming more challenging. "You can tell your kids to get off social media all you want, but it still gets to them."