Twitter is so ‘toxic’ that Amnesty International is enlisting a ‘Troll Patrol’ to help police it

By Katie Canales

Amnesty International is taking it upon itself to help police harassment on Twitter.

On Wednesday, the human rights organisation commenced with Troll Patrol,” a project which will recruit volunteers to find and analyse potentially abusive tweets directed at female journalists, activists and politicians.

In the short term, the idea is to sift for patterns in these abusive or harmful tweets, and publish the findings. Amnesty’s long-term goal is to use these classified tweets to build an algorithm that Twitter could potentially use to automatically detect abusive tweets without human intervention.

“The more evidence we have to show Twitter how large this problem is and how it affects women of different identities in particular — ideally that will just all put pressure on the company to meet our cause and our demands for the #ToxicTwitter campaign and to show that this is an issue that they are going to take concrete action on,” Azmina Dhordia, technology and gender researcher at Amnesty International, told Business Insider.

The #ToxicTwitter campaign she mentions is a report, released two weeks ago, condemning Twitter for not taking users’ claims of harassment seriously. It’s a 77-page study detailing the results of a 16-month long investigation of how women on Twitter experience online abuse.

All of this comes months after Twitter announced new rules to tackle misconduct on the platform, an improvement that Dhordia said is always welcome when working to better police Twitter’s content. But the problem, said Dhordia, is not the lack of rules.

“Much of the problem that women face when they experience abuse on Twitter is that Twitter’s own policies are not always being consistently and adequately enforced,” said Dhordia.

The trouble with Twitter

Amnesty International’s problem with Twitter’s handling of harassment largely stems from a lack of transparency, says Dhordia. She said the organisation has repeatedly urged Twitter to disclose the number of abuse reports it receives, as well as how many content moderators are employed to review content, and details on how Twitter trains them.

“It’s difficult for us to exactly pinpoint what the problem is if they’re saying that they don’t want to tell us that information,” said Dhordia.

For its part, Twitter CEO Jack Dorsey has committed to being more open and transparent, as he works to shake the company’s reputation for being a safe haven for harassers and bullies. In a statement to Business Insider, Vijaya Gadde, legal, policy and trust and safety lead at Twitter, said “We look forward to ongoing constructive engagement with Amnesty International and others to find real, lasting solutions to ensure women are safer and feel safer online.”

Still, while Twitter has a reporting system in place for users to flag instances of harassment, using it often results in no action from the social media platform.

In one infamous case, Twitter user @hannahgais’s reported a video of a man masturbating, sent to her in a private message Twitter’s review team initially found no violation. Only after directly calling out Twitter CEO Jack Dorsey in a public tweet was her claim addressed. See the thread below.

This kind of abuse extends up to the highest levels of power, making the Troll Patrol’s mission more urgent. According to Amnesty International, Black and Asian female MPs in the UK received 35 per cent more abusive tweets than white female members of parliament in a study completed leading up to the 2017 UK election. Just look at some of the messages received by Dawn Butler, a member of Britain’s Parliament.

Decoding the Troll Patrol

Anyone can become a volunteer for the Troll Patrol through an online portal. Those volunteers will get the title of Decoder. At the moment, Amnesty has recruited 557 of a planned 10,000-plus Decoders into the Troll Patrol.

Once assigned, the Decoder will sift through a designated pool of potentially abusive tweets, which currently sits at 501,796, scanning for any signs of abuse or harassment and classifying what they see. Perhaps it’s a violent threat in the form of a photo, or includes a video that uses sexist or racist language.

Amnesty acknowledges that reading abusive tweets can be traumatic for these Decoders, which is why a disclaimer message will pop up reminding the volunteers to take breaks as needed. A similar message will appear after every 10 abusive tweets are flagged and analysed.

There are limitations to what these Decoders can do: While they can analyse, research, and help sift public-facing Twitter data, private messages are off limits to anybody who doesn’t work for the company. Really, as much as anything else, a big goal is just to demonstrate to Twitter that there’s more that it can — and should — do.

“As a company Twitter has a responsibility to respect human rights and we’re trying to show them how women are experiencing violence and abuse on the platform,” says Dhordia.

See Also:
Snapchat just introduced group video calling
Apple-affiliate says ‘cool new Apple products are coming’
Mark Zuckerberg wants an independent council for removing dangerous content

Read the original article at Business Insider