Three months of content moderation for Facebook in Berlin

Re:Three months of content moderation for Facebook in Berlin by ams6110 on Sunday January 28, 2018 @07:47PM (#16254570)

Why do they need to moderate every post. The problem posts will be coming from a minority of accounts. Ban them on first offense, or maybe allow one warning, and people will figure out pretty quickly that Facebook is not the place to post that sort of thing.

[ Parent | on HN ]

Re:Three months of content moderation for Facebook in Berlin by twblalock on Sunday January 28, 2018 @08:10PM (#16254657)

Bans are easy to circumvent, and the worst trolls tend to be technically savvy enough to create new accounts. They'll do it forever. They'll hide behind VPNs and anonymizing proxies if that's what it takes.

Also, what do you do if the hate speech is posted or linked to by a notable public figure, such as the President of the United States or a member of Congress?

[ Parent | on HN ]

Re:Three months of content moderation for Facebook in Berlin by Spivak on Sunday January 28, 2018 @08:25PM (#16254716)

So? The system doesn't have to be perfect, it only needs to make trolling time consuming enough to reduce it to tolerable levels.

> hate speech is posted or linked to by a notable public figure

Here's how I would approach it. Accounts that are tied to real identities have a much higher threshold for bans -- public figure or not. Sans a credible threat of harm to specific individuals they can largely say whatever they would like, hateful or not. When pseudoanonymous accounts get flagged their choice is to attach a publicly visible identity to the account or be banned.

[ Parent | on HN ]

Re:Three months of content moderation for Facebook in Berlin by twblalock on Monday January 29, 2018 @12:28AM (#16255508)

Even a single mistake will end up in the news as evidence that Facebook is "right wing" or "hard left," and perhaps both at the same time.

If you think this is easy, you don't understand the problem.

[ Parent | on HN ]

Re:Three months of content moderation for Facebook in Berlin by nerfhammer on Monday January 29, 2018 @12:18AM (#16255473)

It has to be perfect. People will be enraged if they can find any example of it not working. The human brain only understands examples, not rates. Facebook is now an all-knowing god that cannot make mistakes or not know something.

[ Parent | on HN ]

Re:Three months of content moderation for Facebook in Berlin by ben_w on Sunday January 28, 2018 @08:00PM (#16254610)

1) Because you don’t know who the offenders are until you show their content to a moderator.

2) Because zero-tolerance tends to result in over-blocking, and they already have a problem with banning content that shouldn’t be banned.

[ Parent | on HN ]

Re:Three months of content moderation for Facebook in Berlin by ams6110 on Sunday January 28, 2018 @08:07PM (#16254644)

I think Facebook knows enough about its users to precisely identify the accounts that are likely to be offenders.

[ Parent | on HN ]

Re:Three months of content moderation for Facebook in Berlin by ben_w on Monday January 29, 2018 @02:18AM (#16255901)

Perhaps, but to the extent that it does that’s a combination of moderation and machine learning — exactly what post several layers above was saying is insufficient.

[ Parent | on HN ]