[chiefalchemist]

> "During that six-month period, Facebook removed more than three billion fake accounts - more than ever before.

More than seven million "hate speech" posts were removed, also a record high."

Any sense of context here? For example, seven million might be a raw record high, but of how many total posts would be far more telling.

Same for fake accounts. Is that rate up? Down? Was there a time when FB wasn't as vigilant (perhaps to inflate its user base figures)? Is there a smoking gun for the SEC here?

[bArray]

I suspect a significant number of posts removed for "hate speech" could have been people sharing videos/documents related to the Christchurch terrorist attack. It depends on the definition of a post and whether the BBC interpretted Facebook's message correctly. The number 1.5 million gets thrown around a lot [1]. I also heard the value 4 million, but cannot find the source. If that's true, then I imagine the details of other such attacks were also suppressed.

[1] https://www.engadget.com/2019/04/24/facebook-christchurch-vi...

[wolco]

I know family members who will report posts between each other as hate speech when they are fighting and calling each other names.

[Scoundreller]

And sometimes you see something that is an obvious scam/spam, and you figure it will get reviewed faster if labelled “hate speech”.

Just like needing technical support, so you call sales instead.

Or when you say you’re going to cancel just so you “unlock” the retention deals.

Such is the MBA league game of modern life.

[bArray]

I wondered about cases like this. I don't think Facebook is well suited to dealing with the social nuances of real-life humans - context matters. Mass false flagging must be an equally difficult problem to deal with.

Not that I make a habit of it, but sometimes something really annoys me on Facebook, but there's rarely an option that explains the actual problem. For example the other day I got a friend request from some young woman which I didn't accept (but left the request there encase it turned out to somebody who didn't look like their profile). A few days later the request turns into a young Indian looking man. I think most people can agree that bate and switch is bad behaviour for a social media network, but I couldn't find an option to report it using their limited interface.

[hodgesrm]

Looks as though body counts are back. Not an encouraging sign for getting this mess cleaned up. [0]

[0] https://en.wikipedia.org/wiki/Body_count

[debaserab2]

What do you mean exactly?

[hodgesrm]

> More than seven million "hate speech" posts were removed, also a record high.

This is an example of a statistic designed to show progress without offering insight into the underlying problem, which is the insidious effect social networks have had on civil discourse. Absolute numbers are of little help in judging whether Facebook is actually addressing that issue in a meaningful way. It seems plausible that the main motivation for numbers like these is to fend off legislation to force social networks to take more responsibility for content, which would undermine Facebook's business model.

It's not all that different from the US military using body count numbers as part of propaganda to support continued funding for the Vietnam war. Here's a better reference to the controversy associated with those efforts.

[0] https://en.wikipedia.org/wiki/Vietnam_War_body_count_controv...