- The training Facebook moderators are put through has been revealed in secretly filmed footage, recorded by a reporter for British broadcaster Channel 4.
- The training, carried out by Facebook’s content moderation contractor CPL Resources, reveals how reviewers should deal with material, including child abuse and racism.
- CPL moderators were told not to delete a video of a grown man beating a child, images of violence, and racist memes.
- But Facebook said the training “falls short of the high standards we expect” and has put in place steps to remedy the situation.
The training Facebook moderators are put through has been revealed in hidden camera footage.
A journalist for British broadcaster Channel 4 went undercover as a Facebook moderator and was shown how to deal with extreme content, including child abuse, self-harm, and racism.
During the secretly filmed training, new moderators are walked through various kinds of content and told to delete, ignore, or mark it as disturbing. The latter means it remains on Facebook, but places a restriction on who is able to view the content.
The reporter was employed by CPL Resources, a Dublin-based company to which Facebook outsources content for moderation. Below are some examples CPL gave during the training, which featured in the Channel 4 documentary “Inside Facebook: Secrets of the Social Network.”
Facebook published a blog about the film, in which it said the training fell short of its standards. Monika Bickert, vice president of global policy management, said some of what was shown in Channel 4’s documentary “does not reflect Facebook’s policies or values, and falls short of the high standards we expect.”
Facebook has reviewed training materials at contractors like CPL and provided refresher courses for moderators at the contractor. CPL staff who do not reflect Facebook’s standards, no longer moderate content on the site.
CPL told Channel 4: “Ensuring our trainers and employees are always fully trained and up-to-date on Facebook policy changes is critically important to us, so we’re investigating this matter as a priority, and taking immediate steps with Facebook.”
Footage obtained by Channel 4 showed moderator trainers running through a number of specific examples of how to handle content. In some cases, the decision was clear: The content should be removed. Others were less obvious.
Pictures of topless women – delete
Early in the documentary, trainee moderators were shown pictures of topless women, and instructed to “delete images that depict real adult females with uncovered nipples.”
Violent death – mark as disturbing
Moderators were shown images of graphic violence and told that they “are not the nicest to look at” and instructed to duck out of the room and “grab a cup of water” if any of the images made them feel ill.
They were told that footage of someone dying is “not necessarily going to be a delete.”
The reporter asked why images of people dying aren’t removed, and was told that sometimes they fit within Facebook’s terms – especially if they “raise awareness.” The trainer did not elaborate on this point on the Channel 4 documentary.
Footage of a little boy being beaten by his step-father – leave it up
The trainee moderators were shown footage of a small boy being viciously attacked by what transpired to be his step-father. This was used as an example of what could be left up and “marked as disturbing.”
The trainer says: “We always mark as disturbing child abuse, we never delete it and we never ignore it.”
In Channel 4’s undercover footage, a trainee asks what must be done to escalate the child abuse footage. They are told by a trainer that “there’s nothing we can do really,” unless it’s a live video.
In Channel 4’s documentary Richard Allan, Facebook’s vice president of public policy, said that the video “should have been taken down.” You can watch the full interview with Allan below:
Footage of a man eating live baby rats – ignore
While on the job as a moderator, the undercover reporter reviewed footage of somebody eating live baby rats. He was told to ignore the video because it was for “feeding purposes” and therefore didn’t violate Facebook’s animal cruelty policy.
Footage of two girls fighting – mark as disturbing
Working as a moderator, the undercover reporter came across footage of two teenage girls in a physical fight, which had been shared over 1,000 times. His assessment was that one girl is “definitely more dominant than the other” and results in the other girl getting “battered.” He asked for clarification about whether to delete the video.
A moderator tells him that this video falls under a new policy, and that because the caption on the video is a condemnation of the bullying, it should be marked as disturbing. Otherwise, any physical bullying of minors is a delete.
Images of self-harm – delete “promotion,” leave up “admission”
Trainees were shown images and memes of “suicide and self-harm promotion.” This included an image of healed cuts on someone’s arm accompanied by the text “miss the way it feels.” They were told that this kind of content needs to be deleted.
However, content showing self-harm with no “promotional” context is called “self-harm admission,” and is left up on the site. The trainer said “anything to do with admission we are going to send a checkpoint.” A checkpoint is a message sent to the user, which contains information about mental health support services.
Richard Allan argued that there is a legitimate reason for Facebook to leave images on the platform. “There’s actually a very strong valid interest from that person, if they’re expressing distress, to be able to express their distress to their family and friends through Facebook and then get help,” he said.
Clearly underage users – ignore and don’t send help
The trainees were shown an image of a clearly underage user posting a picture about having an eating disorder, in which case they were told not to send a checkpoint. They are told that Facebook does not action underage accounts unless they specifically admit to being under 13 years old.
“We need to have an admission that the person is underage. If not, we just like pretend that we are blind,” the trainer said.
Facebook said this is not the case. In her blog, policy chief Bickert said: “If someone is reported to us as being under 13, the reviewer will look at the content on their profile (text and photos) to try to ascertain their age.
“If they believe the person is under 13, the account will be put on a hold and the person will not be able to use Facebook until they provide proof of their age.”
Racist hate speech – it depends
The trainees were shown various racist and Islamophobic memes, and told as an example that images calling for “exclusion, death or harm of Muslims” should be removed as “visual hate.”
However, as an example of what to ignore, they were given a meme of a girl having her head held underwater with the caption “when your daughter’s first crush is a little negro boy.” This was because the image, “implies a lot, but to reach actual the violation you have to jump through a lot of hoops to get there.”
Facebook told Channel 4 that the image does, in fact, violate its hate speech policy.
Hate speech aimed at immigrants – ignore
The undercover reporter asked for advice on whether to delete a comment which said “f**k off back to your own countries” under a video with a caption referring to Muslim immigrants.
The trainer told the reporter: “If it just said Muslims then yeah, you’d take this action. But it doesn’t, so it’s actually an ignore.” Immigrants specifically have fewer protections than ethnic groups on the platform, according to a secretly filmed CPL staff member.
Pages with large followings – proceed with caution
A trainer told Channel 4’s undercover reporter that Britain First’s Facebook page, which was taken down in March, had eight or nine violations when five is the theoretical limit, “but obviously they have a lot of followers so they’re generating a lot of revenue for Facebook.”
Pages with a large following are “shielded content,” and can’t be deleted by moderators at CPL. Full-time Facebook employees make the final call after the pages are put into the “shielded review queue.” One such shielded page is that of far-right activist Tommy Robinson, according to Channel 4, which is followed by just over 904,000 people.
Facebook strongly denied profiting from extreme content. Allan told Channel 4: “Shocking content does not make us more money – that’s just a misunderstanding of how the system works.”
NOW WATCH: Tech Insider videos
Business Insider Emails & Alerts
Site highlights each day to your inbox.