Hacker News Reader

>I can tell you back in the day when I worked for Rackshack I never envied the abuse department. They always looked stressed out.

The police normally rotate staff on and off the analyst post every few months.

I was speaking to a detective, saying I was interested in going into digital forensics, and he just plainly said "no, you don't want to do that". He said he knew he'd spent too long on the job when he was walking through a park, saw a father playing with his daughter, and instead of thinking "aww" thought "you sick fucker".

I have, in the course of my career, come across child abuse content. Personally it didn't really bother me - I just didn't think about it and got on with the job of notifying the police, but I can easily see how being exposed to that sort of thing day in day out would really mess with someone's head.

I still remember reading an interview years back with people working at the National Center for Missing and Exploited Children (I think; it was some time ago) whose job was to identify and track exploited children online. The part that's stuck with me was a remark about an infant whose photos kept popping up and didn't match any missing person reports. They gave the child a name at some point, and could only watch as this child literally grew up through child abuse images, year after year, because they didn't have anything to identify him by.

That interview was haunting. I can't imagine how hard it must be, or what kind of toll it takes on you, even if you're able to rotate through to something less stressful after a short period of time.

> Facebook will have 7,500 content reviewers by the end of December, up from 4,500, and it plans to double the number of employees and contractors who handle safety and security issues to 20,000 by the end of 2018.

Seems like the bulk of these roles are contract. "$24/hour" was mentioned in the article. Probably no health benefits. Probably no counseling.

251 working days * 8 hours * $24 = $48,192. Less taxes. Less sick days. Safe to assume these workers aren't all going to work out of the main office in SF.

Seems like Facebook would have to set up some sort of work from home system, like people who transcribe medical records, or some sort of off-shore system, to make this work for workers. But... spreading out the people who do this work is likely to increase security risks that sensitive data will be shared, and reduce workers ability to find comfort with peers.

No way to look at this where automation isn't going to be better.

This definitely seems like a good place for either regulation or unionization.

Unionization would likely be difficult though due to the low barrier of entry for the job - artificial barriers would have to be created by the union and those break down pretty easily due to human factors (see Walmart).

I'm all for regulation in this case. Require that folks who review content for illegal posts be limited in the amount of content they review in a weekday, must have certain benefits like mental health counseling, and get some vacation / sabbatical from the work.

Risks might be reduced innovation in the field, but the big boys like Facebook and Google have already found excellent ways to snuff out their competition within the existing market.

Unionization doesn't work very well when you can trivially outsource the job/task. It'd be necessary to entirely remake the US economy, because you'd have to enable union powers that could directly control all actions by the parent corporation, including setting up overseas / outsourcing.

If those content reviewers are subcontracted out to vendors in the Philippines or India, they'll be earning a lot less than $50k/year. The average call center agent in Manila earns under US$4k.

I am optimistic about human in the middle. I imagine a large grid of small thumbnails and the Google Captcha style "Choose the pictures that are not gruesome" (i.e. differentiating between violence, a video game, and a red scarf). Many reports could be resolved quickly this way. I would be interested in a more proactive solution that is well documented publicly so that we can apply the tech to blogs/news.

> Its not about how much work they are doing, its about the content they are subjected to look at and review.

If you'd like to experience this, you can make an experiment yourself. If you're a Twitter/Facebook user, look for a group/account you know will have unacceptable content (it's easy, search for white genocide or similar - there's always something available). Then try to find some things that are actually reportable without the need to explain the reason in more than a few words.

After you find 10, you'll likely find it's mentally exhausting to deal with it / look at that content anymore. On the up side, you'll likely report and remove some amounts.

Any neural network can probably be gamed by people that really want to get around it.

That might be true. But remember we are talking about people posting in public places like Facebook. I can't imagine that being a smart move, so I am questioning their intelligence.

> I am okay with 100 good post being automatically removed it if means 1 bad post is also removed.

Why? Posts on the internet are pure data. I'm okay with no posts ever being removed.

Posts on the Internet may be pure data, but that doesn't give support for why they should never be removed.

IRL analogy: the dog shit that my neighbor throws onto the sidewalk is pure matter. Just atoms and molecules and such. Should it not be removed?

EDIT: Not that I agree with the 100 sacrificial removals for each good one. I don't. But the "pure data" argument doesn't work for me.

If you support bad posts being removed from the internet, do you also support bad text being removed from people's hard drives? What about bad thoughts being removed from people's minds?

It's an enormous leap in logic to go from things removed from public forums to private hard drives, let alone one's mind. The fact that creating boundaries is a difficult and messy business does not mean they're not worth creating, even with the flaws.

You sound like things are in absolutes, black and white, it’s not.

Would you remove child porn off the internet? How about a rape victim’s photos and videos? If you could delete something that would cause people to commit suicide, would you?

Why remove when you can show only to specific audiences, ie. police etc. that can do something more than delete the content.

You can delete it and share the information with the proper authorities. In fact that is probably what's done.

Your bizarre, logical, absolutist view of the world is not one the majority of people share. There is a difference between removing the sort of terrible posts that show up on Facebook and thoughts in people's heads. Most people can see the difference. If you can't, I suggest you take some time to consider why other people might see a difference.

That sounds like a policy that would efficiently erode trust and cause an exodus from your platform.

The platforms are in a tricky position, but I agree this job should pay more with perhaps some regulatory oversight.

> However on the internet with regards to what normally amounts to pointless shit that people post on the internet. I am okay with 100 good post being automatically removed it if means 1 bad post is also removed.

If you don't like Internet then just get off it!

You probably never noticed, but some people have this thing called self-actualization. By removing good posts you step on it with your greasy foot. By removing controversial posts you also step on it. That's a lot of damage that you did to innocent people right here.

We've got to get off HTTP just that people like you never have the chance to step into other people's homes, decide what they can and can't post and view.

> By removing controversial posts

We're talkng about images of children being raped.

>By removing controversial posts you also step on it.

We're not talking about "controversial posts" here. It's perfectly legal to be a Nazi, even if Facebook might prefer you not be a Nazi on their platform.

This is about child abuse material. In no way is that being illegal "controversial".