A week ago, YouTube and Facebook said they would block people from identifying the government official thought to be the whistleblower who set in motion an impeachment inquiry into Donald Trump.

It hasn’t worked out so well. A name believed by some to be the whistleblower has been shared thousands of times on Facebook. Videos discussing the identity of the whistleblower have been watched by hundreds of thousands of people on YouTube. And images professing to be of the person have circulated on Instagram, which is owned by Facebook, under dozens of hashtags.

The purported name of the whistleblower appeared on Facebook pages that, combined, were followed by over half a million Facebook users, according to CrowdTangle, a tool that analyses interactions across the site. It is unclear how many of those users saw the post, but the name was easily searchable within various Facebook pages, including right-wing news sites and an individual running for congress.

The failure to keep this official’s name off social media is the latest indication of how difficult it is for these companies to police their sprawling platforms. Armies of human content moderators and screening with artificial intelligence have often proved unfit for the task, particularly when many people are intent on breaking the rules.

A mishmash of policies and lack of coordination among the companies have added to the challenge. Facebook and YouTube said they would block attempts to name the whistleblower, but Twitter said naming him would not violate their policies, so long as posts did not also have more personal information, like an address or phone number.

ALSO READ  How to download Instagram photos from any device

“The policies created by the social media companies look great in isolation. They look great on paper. But each time something like this happens, it shows the problems of each company acting in a silo,” said Claire Wardle, executive director of First Draft, an organisation that fights online disinformation. “Companies don’t yet have a way to do this in a way that is effective, and something more than whack-a-mole.”

YouTube and Facebook said they would remove any mentions of the whistleblower’s name from their sites because naming the whistleblower violated their so-called coordinated harm policies, which prohibit content intended to out a “witness, informant or activist.” The companies would not say how many posts they have taken down.

But Facebook users, for example, have been creative in their efforts to sidestep the company’s content moderation. They have avoided using the name within the text of their posts, which could alert artificial intelligence systems screening for it. Instead, they have included it in the URL or inside an image. Others intentionally added characters such as dollar-signs and asterisks to avoid Facebook’s automated moderation.

Twitter, notably, did not block naming the whistleblower. That allowed Donald Trump Jr, who has over 4 million followers, to tweet a name. Numerous conservative commentators spread Mr Trump’s tweet, and began sharing images and videos on social media channels that they believed showed the person’s identity.

YouTube had taken additional steps to make it difficult to search for the name, removing an autocomplete feature that filled in the name once a person began typing it. If, however, the whistleblower’s full name was already known, users could easily find it.

On Tuesday morning, the top three videos claiming to share the whistleblower’s name each had over 100,000 views. The videos often avoided using the name in the title over the video. Instead, they shared it in the comments and discussions below the video, where people also linked to blogs discussing the identity. Many videos with the name were taken down after The New York Times asked YouTube about them.

Facebook was unwilling to comment on whether it was seeing a coordinated effort to spread the whistleblower’s name, but said it would continue to take down content that violated its policies. YouTube did not respond to a request for comment on the videos being shared on its site.

On Instagram, photographs purporting to show the whistleblower with a number of public figures from the Democratic Party were also widely shared. Earlier in the week, Instagram had blocked searches for the last name being circulated, or for the hashtag #whistle-blower. But on Thursday, it was possible to search on that hashtag, as well as the purported full name or variations of the name. Instagram said it temporarily blocked searches for #whistle-blower while it calibrated its systems to ensure that legitimate searches for the term, or the purported name, were not being affected.

A Twitter spokeswoman said posting names and images did not violate Twitter rules. She said Twitter bans only specific types of private information, such as sharing a person’s home address, private phone number or government identification without their consent.

Donald Trump, and his supporters, have called for the name of the whistleblower to be made public. On Wednesday, during the House’s first public hearing in its impeachment inquiry, Republican members of Congress repeated calls for the whistleblower to be identified.

The New York Times and most other major news organisations have not named the whistleblower, though The Times and others reported he was a CIA analyst.

“Here we are, in the midst of impeachment hearings, and we are witnessing the real-time seeding of a storyline by conspiracy theorists through social media,” said Joan Donovan, a research director at Harvard University’s Shorenstein Center.

Sharing the potential identity of the whistleblower also helps bring together people with like-minded views, Ms Donovan said. Communities formed online to spread certain types of content tend to work together again and again, creating organised campaigns aimed at overwhelming social media companies.

Companies do not have the technology yet to take down every piece of banned content as it is uploaded, Ms Wardle said. And when people are able to organise and coordinate on one social media platform to spread information on another, it is nearly impossible to track, she said.

“People are testing the platforms ahead of 2020. This is a perfect moment to see what works, and what doesn’t,” she said. “So far, we know what doesn’t.”

The New York Times



READ SOURCE

LEAVE A REPLY

Please enter your comment!
Please enter your name here