Twitter’s New Anti-Harassment Tools Are Good, But Still Not Enough

Tuesday morning, Twitter released a new batch of tools designed to combat its congenital harassment problem. In an evasive statement, Twitter blames abuse, bullying, and harassment on a wider trend "across the Internet." The company says it has faced "challenges keeping up," and promises it is "going to keep listening."

The slippery tone of the statement does little to address the public frustration at Twitter's long record of inaction. Fortunately, the tools themselves look solid.

Users will now be able to mute specific keywords from their notifications, as well as mute conversations. If user with a large number of followers sends harassers your way, you can eliminate the notifications with a few clicks. It's a positive step forward—if a very minimal one.

Unfortunately, I have the equivalent of 7 PhDs in harassment on Twitter. As one of the primary targets of Gamergate, I've had hundreds and hundreds of threats to my life on Twitter's platform. Newsweek even did a study showing I was the most harassed figure during the height of Gamergate. I've had diagrams of my home posted on Twitter, along with pictures of my car and license plate, as well as photos of my pets.

Twitter's new solutions work by hiding content from you, rather than removing it. This is a solution that can leave you in more danger. If someone is threatening to kill you, or if private information about you is released, you're more likely to simply be unaware of it. One of the reasons I don't use block lists (a previous Twitter safety update), is that I need to be aware of threats on my life and my audience when I speak publicly. Similarly, if someone is spreading private information about you, or releasing revenge pornography, with keywords you've blocked or in threads you've muted, you just won't know.

Yet I can't deny that similar measures have made Twitter much more pleasant to use. In August, Twitter released its "Quality Filter" to everyone and turned it on by default. This is the single most effective action Twitter has taken, and it hasn't gotten credit for it. Instantly, the amount of harassment I received from bots plummeted to almost nothing. And the alt-right trolls running multiple accounts with just a few followers have completely disappeared. Twitter has erased small-scale harassment, and the company's new tools could erase larger scale harassment.

In many ways, Twitter is held to unfair double standards that other social networks are not. Twitter is actually working on the problem. I've asked Reddit dozens of times over the last few years to remove various threats and doxes, ranging from private medical information to education records obtained by impersonating me. I've had Reddit ignore requests for help when men have sent videos to me holding a knife up to a camera, promising it would be the last thing I'd ever feel. Twitter cares some, Reddit doesn't care at all.

Facebook is even worse, frequently doing nothing over various death threats and doxes. When Facebook shrugs at a violent threat on your life, you're given an option to get support—by messaging one of your friends. There are groups set up to dox and harass transgender women on Facebook, and when you report them nothing happens. My efforts to backchannel with Facebook about this policy have previously gone unanswered‚ even at last year's high profile harassment summit at SXSW. To its credit, Twitter is at least making an effort to curb hate speech towards transgender people, training its staff how to respond.

(Update 11/17: Responding to this critique, Facebook reached out to me to clarify that its rules do not allow harassment based on gender identity, and updated its public documentation to reflect this. However, the core issue isn't a web page, it's Facebook's record of enforcement. It's my hope that, like Twitter, Facebook will train its moderators to better recognize transphobic harassment.)

The problem is, we've heard all of this before. In February of this year, Twitter announced its "Trust and Safety Council," filled with academic experts from groups commonly targeted on Twitter. After the announcement, I expected things to improve. A statistical analysis done by my office showed that things got drastically worse. In the last six months, Twitter's success rate at responding to doxes has plummeted from 63 percent at their height to 0 percent for the last two months. I've personally stopped reporting them, because there is simply no point.

The tech industry has a strong bias towards technical solutions to social problems. In November of last year, the only black member of Twitter's engineering leadership quit, forgoing a severance package to tell his harrowing story on Medium. He describes a meeting with Twitter leadership when they asked him to build a hiring tool identifying people with last names that seem black, ignoring that these tools have a record of missing many candidates. I am white, but with this type of tool, I would be classified as Asian. He asked himself how he could continue to work with people that had such a cognitive blind spot about race, and chose to leave instead.

Twitter's financial turmoil casts doubt if the company will be able to hire enough staff to deal with its problems. Maybe today's announcement indicates that Twitter will start taking doxes seriously again, but I doubt it. I expect it to double down on tech gimmicks like the ethnic-name analyzer.

The main thing Twitter needs to focus on are implementing its rules more uniformly. If outing a transgender woman is against Twitter's rules, that needs to be implemented every time. As of now, with death threats, doxing and spreading revenge porn, it's a crap shoot—your chances of action are somewhere between 0 to 63 percent. The company also needs to improve its ability to respond to harassment spam bots. As of today, someone can pay as little as $5, and send hundred of automated users attacking you and everyone you interact with.

We should all want Twitter to fix this, because it's one of the greatest engines for social change ever invented. The mainstream media ignored police violence in Ferguson until it trended every night on Twitter with pictures and video of protesters being beaten and tear gassed. For all the disasters of 2016, Twitter has brought two major features to its service addressing harassment. If the company is going to survive, these kinds of incremental improvements must continue.