Why Blocking Trolls Doesn’t Work

During the Democratic Convention, journalist Jose Antonio Vargas tweeted a fact:

Facts should not be controversial. But Vargas is a Pulitzer winner who famously outed himself as undocumented in the New York Times, and some of his 70,000 Twitter followers, perhaps aware of this, grew angry nonetheless:

“If I had that deep an animus for the country in which I was living as a guest I would not stay, but that’s just me,” wrote one.

“Wow, you’d never pass the citizenship test,” wrote another.

“Freaking hypocrite! White men didn’t start slavery. But we ended it in western civ. #yourewelcome”

And so on.

All this for a pithy observation about the smart, albeit very flawed, guys who wrote a bunch of this country’s founding documents. But the blowback was not an aberration. Even when you post something innocuous on Twitter today, harassment is “just kind of what happens,” Vargas told me recently.

On Twitter, there are limited tools for dealing with abuse. You can mute people (they still follow you, but you can’t see them) or block them (you bar them from seeing your account). That’s about it. It’s a tremendous frustration. So much so that I have come to think that Twitter founder Jack Dorsey’s greatest achievement is the creation of the world’s most efficient platform for bigotry, white supremacy and harassment.

That’s a painful realization for me. I love Twitter. It’s changed my life, mostly for the better. I’ve made dear friends there. But it’s increasingly obvious that, for those of us who aren’t white or male—I am both—it’s often a miserable place.

Every so often, when a celebrity gets hassled, this makes news. Most recently, Milo Yiannopoulos, a conservative journalist, sicced his followers on Ghostbusters star Leslie Jones, because Jones is black and a woman. The abuse—one popular tactic was to compare her to either an ape or gorilla—got so unbearable that Jones, for a time, left Twitter. (After to many people’s surprise Twitter banned Yiannopoulos, Jones returned; incredibly, her commentary on the Olympics proved so popular that she’s now joined NBC in Rio.)

A study published earlier this year captured the scope of the ugliness:

A couple years ago, I made the case for blocking on Twitter as a worthy solution to this problem. But in hindsight I was both optimistic and myopic. I can block people on a whim with no thought to consequences. My mentions will not turn into a cesspool. I won’t have to spend my day justifying my existence. But that’s precisely what Twitter, in enabling the abuse—or, at best, doing very little to stop it—has forced people to do.

As Anil Dash, a writer and veteran of the tech industry who has been Twitter since just after it launched about a decade ago, told me, “All you’re ever doing is petitioning for your humanity.”

“I’m just some dude on the Internet, and somebody will tell me probably once or twice a day to kill myself,” Dash says. “I think if I look back to my late teens, early twenties when I was struggling with depression—if I had endured somebody telling me once or twice a day to kill myself, as happens now—it would have worked.”

Twitter’s anti-abuse tools often are not strong enough to keep even a handful of trolls at bay, let alone a phalanx. This has forced users to find oftentimes maddeningly creative workarounds, if they choose to stay at all.

Last year, for example, more than a hundred white supremacists coordinated the harassment of Celeste Ng, author of Everything I Never Told You. Despite utilizing third-party troll-blockers like Block Together and Twitter Block Chain, a Google Chrome plug-in, Ng routinely deals with racist and misogynistic tweets, particularly when she posts about Asian representation in Hollywood or anything Trump-related. Before she installed the software, she dealt with five to 25 abusive tweets a day. That, she said, “doesn’t sound like a lot, but opening up your mentions every morning to people saying hateful things just gets kind of wearing.”

Initially, she blocked and reported harassers, or replied to a few tweets to “model some ways to respond.” And that was fairly effective. The problem, however, “is that the really dedicated trolls love to be blocked. They take great delight in creating a new account to troll you again, and in tagging in their troll friends to come and harass you, too. It’s a hydra effect: You block one, you get half a dozen new ones.

With the white-supremacist infestation, Ng had had enough. She decided on a new tactic. She began “donating $10 to the ACLU for every harassing comment.” In a couple of cases, her antagonists were doubly punished. “Based on your comment, profile pic & feed, I’ll donate to both the ACLU & Planned Parenthood on your behalf,” she informed one.

All told, Ng donated about $200. “Small price to pay,” she told me.

For more on trolling’s effect on the Internet, read TIME’s new cover story.

Sometimes a more proactive approach is required.

Aminatou Sow, a digital strategist in the tech space, tweeted in July that a high school–age boy was “talking nonsense about diversity/race” to her. She identified him, contacted his school and cc’d his parents. On another occasion, a man was overtly racist. “I called up his work and I was like, Hello. Here’s what your employee is doing while he’s representing you in a public space,” she recalled to me.

This was cathartic, Sow said, “but unfortunately that’s not a mechanism that you can scale and replicate every day.”

Sow doesn’t block often, particularly if she knows the assailant personally or professionally. Instead, she favors muting, because that leaves the harasser screaming into the void, which she loves. But either way, the available tools, she says, are insufficient: “I’ve blocked people and still seen their garbage, or I’ve muted them and still have been forced to interact with them.”

Sow thinks about leaving Twitter all the time. Not long ago, over the course of a single day, three different people called her a nigger— more abuse in 24 hours than I have in years of being on Twitter. “At the same time, I’m just very conscious of the fact that what these people do want is for us to want to leave Twitter. It’s like, why should we leave it to them? That is just not fair.”

I hope there’s a silver bullet—an as-yet-undeveloped troll-stopping tool—but it has so far eluded both users and former staffers alike. Blocking cannot stop people who tweet racial epithets or demands that people end their lives. Nor can muting. Nor would an algorithm designed to target keywords.

Twitter’s original sin, perhaps, was not building in safety and security tools from day one. But even then, I cannot imagine how one could function.

Twitter is too large—an international company with 140 million active users, 79% of whom do not reside in the United States. The number of abuse complaints Twitter receives would, I expect, boggle the mind. An actual person must review them; it’s not something that you can automate. And because it’s an international platform, these people must be well versed in a vast array of racism in order to recognize it. It does not seem feasible for Twitter to hire enough people to understand the nuances of potentially hundreds of forms of ever-changing harassment.

As Anil Dash noted, while Americans are upset about Leslie Jones (and most of the people who do get help have either very public profiles or advocates, or internal connections), Twitter must grapple with people, apparently from the Middle East, who “report[ed] ‘atheist’ and pro-LGBT girls and women to the local authorities in places where blasphemy laws allow for punishments as severe as death.”

“It’s a no-brainer what you should prioritize,” he says. “But those things are kind of invisible to one another because they’re different communities.”

While it’s not as if a large social media platform cannot deal with abuse—Facebook is famously more adept—Dash points out that not being required to follow people back on Twitter means we aren’t beholden to each other in a way we are on Facebook. And without asymmetrical following and Twitter’s hashtag infrastructure, you also wouldn’t have the Black Lives Matter movement, which spread like wildfire on Twitter but has been considerably less effective on Facebook. What makes Twitter terrible makes it great.

Jack Dorsey has said that online harassment “has no place on Twitter.” But that rings hollow, and sounds a lot like politicians who, in the wake of some atrocity, say that America “is better than this.” But maybe white supremacy is, in fact, what Twitter—and America—is good at.

Judging by its actions, Twitter seems to believe that’s just fine. It seems more focused on growing the platform, adding more advertising tools and pleasing publishers, at the expense of its most active users. Perhaps the harassment isn’t losing enough users to offset the cost of implementation. As Sow put it, “I’ve seen people leave Twitter, who always said that at the end of the day they feel like those useless egg avatars are more important to Twitter’s growth than they are.”

And that sucks. Celeste Ng shouldn’t have to wade through garbage. And yet she persists. “Trolls want to silence you,” she said. “I don’t like being intimidated, so I figure the best way I can counter this is to keep talking about the things that are important to me, trolls be damned.”

Jose Vargas is similarly stubborn. The chance that even his harmless tweets will piss people off, he says, is no reason not to speak his mind: “If I worried about what people were going to say, then I probably wouldn’t be tweeting anything at all.”

I find Vargas and Ng’s bullheadedness in the face of frequent harassment extraordinary. And I hope Twitter, somehow, does right by them—allows them to tweet without fear of reprisal. As an undocumented immigrant who once made passage from Great Britain to Philadelphia would have it, that’s just common sense.