Open Letter from Facebook Content Moderators
foxglove.org.uk
foxglove.org.uk
Low pay and no psychiatric help with that kind of content?! This is just outrageous. Mark should moderate this stuff by himself to see how huge burden it is. And, this is not Rotten or Goatse type of shock it’s entirely different level.
Then let them say in public that they're doing their best for their employees.
Mining magnates should be forced to work at the coalface. The Waltons should be forced to stock shelves, Bezos should be in the distribution centres.
The list is endless, and unfortunately this is nothing new. The best we can do is minimum wage laws and unions to try and "negotiate" with billionaires to give us a crumb.
People only have so much time and capacities. A software engineer isn't going to spend a lot of time on the sales room floor, but they might have started off in help-desk support. A CEO might not know what goes on in the warehouse, but he or she may have worked at a McDonalds growing up. If someone hasn't worked entry level jobs elsewhere, would working on the factory really develop empathy? Would that change policy? Probably not.
It's so easy to make two boxes, one for the oppressor and the other for the oppressed, and place everyone into those two boxes. If each oppressor works a day in the shoes of the other box, does that make a difference? If the workers throw their wrenches into the gears and seize the means of production, what if they're completely unprepared to keep that machine moving and end up with famine?
The issues are complex, much more complex than "Before you get your license, you should have to ride a bike on the streets for two weeks."
You've got a point there. The capitalist will most likely continue to act as a capitalist. The incentives are fundamentally wrong, so moralizing on an individual level won't do much, which brings us back to: blame capitalism.
To say he has not been made aware and had the opportunity to address these issues over years is giving him too much credit.
Blame capitalism is all the rave. So are companies hiding from responsibility. Shell companies shell companies to evade responsibility and incing their companies on remote islands they clearly don't primarily operate on to evade taxes and giving abuses NDAs which effectively offer full silence regardless of any federal felonies committed or else they take severence or healthcare taken away. Companies can do whatever they want under an NDA and sue anyone to the point even if they did something very illegal, they still wouldn't be liable because the fact that someone spoke about it is illegal in the same way a VC investor was caught beating a woman 300 times in a row as she tried to escape the guys san Francisco mansion but the evidence was thrown out because it was obtained illegally and he got a $500 fine instead.
Blame capitalism exists because publicly shaming companies away from egregious abuses even to the point of being often illegal is the only way to have some type of check and balance and it always takes a few people beginning by speaking out and losing their jobs and getting nothing for it before the masses come to agreement and hold companies responsible for better practices. Unfortunately this is not just a fun trend. It's a reaction that attempts to check and balance abuses and it often only happens after many yrs and many people have gone silent and a few have chosen to take the risk and speak out. We should definitely have something better than blame capitalism but to say it just appeared out of the blue is a bit misleading.
It tends to encourage it by putting those at the top who manage to exploit the most. Capital also begets political power which makes it hard to counter that trend.
> Wanting more and more money regardless of the (human) cost
Congratulations, you've accidentally defined capitalism.
What I'm advocating against is the idea of a wholesale changing of a system of politics and economy.
Some countries do better than others, in various metrics. That's the "what else" you can look at and emulate.
True
> It doesn't even really encourage it.
False
> Wanting more and more money regardless of the (human) cost does.
Capitalism is explicitly about encouraging a behavior with more and more money.
It's not immoral per se, it's just that humanity is not in the value function. That's where regulation/subsidy/taxing comes in.
Falling back to social engineering and forced labor (hopefully) won't win you many points in the "let's re-architect the system" debate. If the Waltons never stopped stocking shelves and started scaling, we'd never have Walmart. It's possible you think that's a good decision for the world, but it's clearly not the worlds decision for itself.
If there's one billionaire with intimate knowledge of the "lower" levels of his business, it's Bezos.
I'm not here to tell him how to run his business, but let's not pretend he would be happy to work in his warehouse.
I’m a computer guy, but there I was loading trucks, packing boxes, etc.
It was a fantastic experience and connected you viscerally to the business. I believe it is still required. This is the culture that Jeff shaped for the company.
I wouldn't overstate it though -- the ability to tolerate a week of something does not necessarily extrapolate to a month or a year.
A week in the rainforest is an adventure. A life in the rainforest is difficult and short.
But I don't intend any criticism of Bezos. He has done the math on warehouse employee satisfaction as a contributor to corporate goals, and made his decision.
The scrappy days of packing his own shipments are long past, and that experience gives him not-much insight into the working conditions for the people in his warehouses.
That's fine, but it's not a basis for deification.
For you, I imagine it was a week-long challenge where you knew that you would have a better job afterwards and make a lot more money.
"On the first day he took me to the officer’s bathroom and told me I would be responsible for keeping it clean. 'Here’s how you clean a toilet,' he said.
"And he got down on his knees in front of the porcelain bowl, in his pressed starched spotless dress uniform, and scrubbed the toilet with his bare hands.
"To a 19 year old who has to clean toilets, something which is almost by definition the worst possible job in the world, the sight of this high ranking, 38 year old, immaculate, manicured, pampered discipline officer cleaning a toilet completely reset my attitude. If he can clean a toilet, I can clean a toilet. There’s nothing wrong with cleaning toilets. My loyalty and inspiration from that moment on were unflagging. That’s leadership."
I did just that on my last project and it was great. Another team was developing a similar product, but they didn't have an 'in house real user' so theirs was much harder to use.
When I ran a development shop, every year the senior staff spent a week physically working in their client's department eating their own dog-food (i.e. using the apps they built/supported). Once this was implemented it cut down on the post-implementation change requests by over half.
1. Limited provable return on investment. The effects are real but it's not like they're a panacea.
2. Human nature. There are a lot of people who are not enthusiastic about participating — because they are not cut out for the positions they're exchanging into for a spell, because they don't see the point, because they occupy a privilege place in the office hierarchy, etc.
2b) if they don't see the point, there needs to be a serious conversation and maybe consider a new hire who does understand the benefits
2c) yeah, if you're too important to want to improve the business, let's have you talk to the two Bobs. What is it you actually do here?
In my experience advocating for such initiatives — and I have done so across multiple organizations throughout my decades in the workforce, and not just at tech companies — it is not so much resistance from below as from above that stymies them.
I don't have a ton of experience with SoX compliance specifically.
The first two days you spent replying to customers, and the third day you spent prototyping or writing up some way that you thought common complaints could be addressed by making changes to the product.
It was a nearly unmitigated boon to the company. Sure, the company saw it as paying for a few extra vacation days. But the program stuck around because it generated a lot of good ideas, and the direct communication between CS and product development teams was valuable.
But it was worth it.
Quickly pops people out of their focus on local optima, to better see the whole.
A couple of days is not enough to get the gravity of the situation. Spend a week or more looking at the filth that needs filtering and you might start to get a sense of what the mods deal with and how badly it can affect them. Maybe.
I've always thought that was a fantastic form of training, and spoke strongly to the fact that the chain has been around since WW II.
It's possible you were right, if we can interpret the "attack" as being purely intended to undermine the original argument.
I just didn't personally interpret it that way... it looked more like a side-observation to me, something like "name-calling" under Paul Graham's categorisation:
https://en.wikipedia.org/wiki/Paul_Graham_(programmer)#Graha...
I see no reason why there can't be a mandate to provide mental health benefits to these workers. It is absolutely a safety hazard.
Individual workers might not price in the societal (or individual) cost 20 years from now of breeding mental issues in these workers now.
Or another example: You own 50% of the forest land in your community, which for sake of argument produces 50% of the community's oxygen (assuming a closed system). You start cutting down the trees to produce furniture and sell them. Generally the cost of oxygen production is not priced in, and down the line, someone will pay for it. Either individuals getting sick, universal health care, environmental efforts in replanting, regulation on land maintenance, etc, etc.
I suppose you will argue that each individual will need to save a buffer for every network effect for every purchase decision they ever make that doesn't price in externalities, in pure capitalist spirit. But I hope my point hits that there has never existed such a system, or if it did, it died out pretty quickly, and left no trace in history.
We can and should all have opinions on health regulation, environmental taxes, etc regarding what their effect ultimately becomes, but everything from societal mental health to environmental effects will always leave a bill, and someone will pay for it. The only thing we can discuss is if it should be everyone for themselves, our great great grandchildren, taxes, or corporations selling crap too cheap (or in OPs case for free).
I spent ~6 months working on tech to detect potentially-suicidal people online (which required a lot of manual flagging and review) and it's one of the most emotionally draining jobs I've ever done. It's a weird feeling to think it's an incredibly important job, but not want to recommend anyone ever do it -- that seems like a perfect role for automation to slip in.
I have no faith in AI to handle nuance like this: https://www.theregister.com/2020/10/28/gpt3_medical_chatbot_...
The Rotten Library was actually a fascinating place, though: https://www.gwern.net/docs/rotten.com/library/whatsnew/index...
I’d expect almost everyone who works at a computer screen to be working from home since March. Even if it requires expensive technical solutions to ensure it. I can’t imagine moderators need more than an internet connection and computer.
The whole thing in the article about Facebook suddenly trying to automate this part of the content moderation job seems to also point in that direction.
> Management told moderators that we should no longer see certain varieties of toxic content coming up in the review tool from which we work— such as graphic violence or child abuse, for example.
Given that no other facebook employees are being forced to work in their expansive offices perhaps the content moderators, if they are required to work on site for security purposes, should be given the run of the empty space. That way they can easily social distance appropriately.
My principle concern would be, no matter the employment relationship, you can't contract away criminal liability. (At least, not in the US.) So here you opened something while working from home, that maybe your state's child sex trafficking task force was tracking. It could just get messy. I'm not sure I would even want to do that kind of work from home.
In healthcare, totally different but highly regulated, we can let people work from home with HIPAA protected info by securing the way they work and access the info.
There must be some solution though, like requiring that the VPN connection originate from an address considered to be in the same state.
It seems ridiculous, but very possible.
That doesn't excuse Facebook's despicable treatment of their contractors who have to wade through that filth.
If the video feed goes dark, the surveillance operator could trigger a forced shutdown of their laptop (or even better, disconnect their RDP/VDI/etc. session).
People are very psychologically affected when they feel they are being constantly observed, especially when that observation is only for punishment.
There's a story from an Amazon warehouse recently where a guy had a heart attack and didn't get found for 20 minutes. A few days prior he had put something back in the wrong place and management came to him in 2 minutes. Do you think Facebook overseers would be any different?
Within two weeks all employees, which weren't not needed at the front line or personally in an office were moved to home office. The company was able to move 10s of thousands of employees in a couple weeks securely and reliably (safe for a few glitches, which are to be expected on that scale).
There are reliable solutions to completely segregate what's happening on corporate systems from a simple, employee owned laptop connected from home.
Sure, you can make pictures of the screen, but you can do that in the office too. Then again you can just print it, while at the office. Something which you can't do at home.
So no, I don't really buy this reasoning for a company fancying itself totally hi-tech.
You would have to pay me far, far more to watch animal/child abuse professionally than to write JavaScript.
How is this a low-paid job?
People choose these jobs because they are the best options available to them. This is true for many undesirable occupations (such as janitorial services).
It's kind of amazing what unions can achieve when they're allowed to exist. I think cops are the only folks who get to unionize without the cops beating them up.
Yeah, nurses and EMTs should be paid more. All the EMTs that serve Seattle get "based" out of city limits and don't even make the Seattle minimum wage. You can make more as a dishwasher, and dishwashers don't have 24-hour shifts.
While corporations like FaceBook would purport to have liberal values, in reality they hold these beliefs only insofar as it furthers their own financial well-being. While the elites in this company seek to concentrate wealth and power, those at the bottom receive scorn and poor working conditions. Image having to work such an emotionally draining job, and not being given the mere dignity of being a full-time employee.
This sounds rather dehumanizing in my view. You are just there for labor, and they will do everything they can to minimize the amount of monetary value that they provide to you. This means humane working conditions, inclusion, respect, and psychological support and are all liabilities for them.
Would they vote en masse to form a Union to stop these issues arising in the first place? Personally I doubt it.
It seems to me that many people are happy to take the large pay check every day, and see themselves as "the good guys" (as everyone does) because they donated to Biden and attended a BLM march or put pronouns in their bio.
Yet the homeless problem outside their front door never seems to get better, and they have terribly paid and horribly treated people working in their own company, and they seem (from the outside at least) to be doing very little about it.
Of course Facebook is just one of many companies with the same problems. And the political Left abandoned the working class a long time ago. It's a systemic issue.
Where else can an employee browse filth all day in a secured environment? Free from the eyes of children who may be doing pandemic schooling from home? And indemnified against criminal liability for opening up certain troubling related content by an authoritative, audit ready, third party recorded video feed?
There are just legal requirements around certain types of porn for example. Or videos of illegal activities up to and sometimes even including murder.
Law enforcement authorities only tolerate a certain number of mistakes in this regard. With certain content, they may not tolerate any mistakes. I'm not saying FB is right. In fact, FB probably didn't even make the decision. They outsource. What I am saying is that FB, with the content they want their contractors to moderate, are putting those third parties in highly precarious legal positions. I don't blame those companies for acting to protect themselves if FB is not willing to take on all legal responsibility for things that may go wrong.
If I understand your point, it is law enforcement doesn't allow moderation of these items outside of a certain environment? If so, can you suggest further reading of the laws for this area? I have never heard of them.
I'm saying as a contractor, as the low man on the totem pole, there is no way I would want any of that content touching my home network without an iron-clad assurance from every level of law enforcement that I would not be prosecuted. An assurance from your local prosecutor probably means nothing to the guys at the US Attorney's office. (And sometimes even vice-versa.) Accessing those materials in a secure environment that is audited and recorded, in direct partnership with every level of law enforcement, avoids issues of that content touching your router altogether.
I can't imagine that any person who understands criminal liability would actually want to do this kind of work from their own home networks.
https://thenextweb.com/microsoft/2017/01/12/microsoft-sued-b...
I'm not sure exactly how this works in tech. I know lawyers who have worked on child abuse cases and they had tight restrictions on the evidence. Only the trial lawyers could request access. No paralegals, no sectaries, no copies. Only select Lawyers, the jury and the judge could view the evidence, and it was all controlled by the Justice Department. (I knew a lawyer who refused to look at the evidence, only the descriptions, because he didn't think he could handle defending him otherwise. He did successfully defend him though).
I suspect for Microsoft, they have some preliminary hashing algos that automatically send known media to a particular people with the same Justice Dept exceptions and controls for chain of custody. Other people may see content of course, but it would likely get flagged and shipped to people who were authorized to deal with it. You don't want a lot of people on that list obviously, but when it's just a few people, they get a constant stream of nightmare fuel.
Imagine the chaos someone could cause.
It'd be like back in 2010ish when somebody encoded a bunch of malware into bitcoin transactions: every bitcoin node running any kind of antivirus software was immediately kicked offline. Many AV programs deleted their owners' copy of bitcoind. It was madness. Forced the bitcoin developers to store the blockchain on-disk xored with a different random number unique to each computer.
They haven't been, just the people contracted by CPL/Accenture to work at FB, which leads me to believe that this is not a legal requirement.
It's presumably that FB said you need to ensure security to the companies, and they decided that the easiest way was to bring everyone back to the office.
If I had to guess, I suspect that the motivation to bring people back to the office comes from content security issues. Not all content flagged for moderators is actually objectionable due to errant or malicious flagging. It's impossible to guarantee that private material will stay secure when it leaves the premises.
I'm on the side of the content moderators here (should be allowed to work from home) but it's not difficult to imagine the headlines criticizing Facebook for allowing private content to be viewed, and potentially shared, by content unsupervised content moderators working at home.
https://battlepenguin.com/politics/is-meaningful-section-230...
The AI wasn’t up to the job. Important speech got swept into the maw of the Facebook filter—and risky content, like self-harm, stayed up.
The lesson is clear. Facebook’s algorithms are years away from achieving the necessary level of sophistication to moderate content automatically. They may never get there.
Most technical people have guessed that Facebooks AI wasn't actually any good. The non-technical world, like politicians, assumed that it was easy for Facebook to moderate. Facebooks magical AI could censor anything that governments deemed offensive. It turns out it can't, but Facebooks has done nothing do inform the world that their AI is no good. Admittedly I guessed that it didn't actually exists and was simply a keyword filter.
I really gotta cancel my Facebook account. What greed.
How many people do you need to moderate the interactions of half the world's population? It's going to require a lot more than tens of thousands of people, and those people will need to be paid a lot more than 18 dollars an hour.
It's not immediately clear from the letter that this was Facebook's decision. In fact, it appears that Facebook subcontracts content moderation to Accenture and CPL, both of which are named in this letter.
The letter also takes shots at Mark Zuckerberg's growing wealth and attacks Facebook's attempts to augment the human content moderators with AI pre-filters to remove the worst content. The latter claim is strange, as the content moderators are the ones most likely to benefit from AI pre-filtering. They try to claim that Facebook "failed" to create a workable AI, but I seriously doubt that Facebook has given up on the project. More likely that they're still iterating. I suspect these people feel they need to attack the AI angle to protect their jobs.
The letter goes on to demand 50% higher hazard pay, demands that the subcontracted content moderators be hired directly into Facebook as full employees, out of their contracting jobs under consulting firms.
Do you think this in general, or just for content moderation? For example, is it reasonable that I can buy car insurance that covers me if I cause an accident?
did you mean to phrase this differently? isn't that the primary purpose of having car insurance?
Reasonable? I’d describe it as mandatory, and not even a question in no-fault jurisdictions. It doesn’t exempt you from criminal charges if you did it, say, intentionally.
That's the right thing to do, considering FB is the only entity with some power over their employers. Avoiding this kind of thing is exactly why brands outsource their dirty work.
Content at the top, buttons down below to classify why it needs to be banned or if it's ok.
If more context is needed i'd still be hiding user information from the moderators but maybe pulling up some number of posts 'around' the suspect post.
Like anyone serving in the military and possibly subject to all the PTSD and injury that combat can entail, these moderators are standing as a thin line against the worst things humanity can do to each other.
They should be given respect and every possible consideration for their sacrifice.
Their demands are merely a simple first step closer to the recognition they actually deserve. I would go much further.
On the one hand, they probably haven't. On the other hand, they provably aren't as close as they thought they were, and may be very far away.
The fact that they haven't fired all the moderators yet should give you an idea of how difficult it is.
How many human moderators does Facebook employ? Zero? AI lives up to the hype!
It reminds me of the filters on facebook that kept killing compliant nude pictures of larger people, because the model had obviously just learned to count pixels of skin.
It'll be a long time before this work gets automated, and until then, they should support their moderators.
That makes them fairly important, in my eyes.
If people had any idea of what crap these people prevent them from seeing, then they would realise the value.
In general, the treatment of content moderators in tech companies is a deep stain on any professed goodness of these companies.
I'm trying to quit IG so my next phone will be fb free but does anyone have any recommendations on how to make dropping IG easier? So far I made a group with friends so we can spam pics but it's not the same (probably cause not enough of us is spamming it)
Does anyone know of a IG replacement?
Whatsapp is the primary way I do this because of it's global reach, but you could get away with sms too.
We should be thankful for the jobs that we have today. Corporations of the future will not need as many warm bodies. (Sarcasm)
So in regions with reasonably healthy people, its not clear that COVID is actually more deadly than the regular seasonal viruses we have been dealing with for the past few decades.
https://www.statista.com/statistics/525353/sweden-number-of-...
https://swprs.org/wp-content/uploads/2020/10/sweden-monthly-...
https://www.acpjournals.org/doi/10.7326/M20-6817
..may still prevent transmission based on the new CDC guidelines, but I don't think they based those on any actual studies.
Honestly the rules are all dumb, random and arbitrary at this point. People keep yelling "science" when it's really just new superstition.
When cancer detection rates go down, it doesn't mean cancers went down, it means people aren't coming in for their regular mammograms and colonoscopies. 5 months can mean the different between an early treatable cancer with a 5+ year survival rate and a stage 3 malignancy with the survival rate dropping by 1/3.
Five years from now, we'll look back and be able to see the massive secondary deaths from cancer, heart disease/obesity (people stuck inside all day), depression/suicide and others. This will not be a time of honor.
> Across the world, rates for attempted suicide are growing dangerously amid the continued lockdowns
This is incorrect. Realtime monitoring data have not shown an increase yet in deaths by suicide. These are early figures, and we do need to guard against rises in future. But there has not been an increase.
https://www.bmj.com/content/371/bmj.m4352.full?ijkey=NDYZgh6...
> During the study period (3 April to 2 June 2020)
This is the period where almost all of Europe saw a decline in daily-new-cases, mainly due to people starting to take this thing seriously.
And still, the study has shown that masked worked:
> The primary outcome occurred in 42 participants (1.8%) in the mask group and 53 (2.1%) in the control group.
What we do know for sure is that masks can prevent a certain amount of exhalation droplets, and there are plenty of other studies that point to their efficacy [0]. Given reasonable evidence in one direction, less evidence in the other, but perhaps still a lack of consensus, it seems like taking a very simple step that could save lives is still the prudent & responsible step to take.
[0] https://www.medrxiv.org/content/10.1101/2020.05.22.20109231v...
Other decisions on things like which businesses are able to stay open, which closed etc., do seem arbitrary at points. However, if lines have to be drawn somewhere, edge cases close to those lines may always seem inequitable. Personally my main criticism is that we don't have stronger guidelines that deal with more detailed scenarios at the Fed level, that states left to their own devices are often no more consistent, and municipalities are often left to try & figure it out themselves.
Your interpretation of the data is not correct.
[0]https://emanuelkarlsten.se/number-of-deaths-in-sweden-during...