Google order its secretive "raters'" hours cut, so now they're going public

Originally published at:


mmmm, a complex, automated standardized testing system* that uses algorithms to accept or reject a human being who relies on them to stay alive…all in the name of training them to train the robots who will eventually replace them. In a company that would already run every service in the world as a totally monopoly/monarchy if left to it’s own devices (pun intended) Do dystopian Sci-Fi writers and Onion writers get together weekly to grumble about how there’s nothing left to write about anymore?

*maybe public schools really are preparing us well for the future… :sweat:


Don’t Be Evil


Balance your work and life, join the Leapforce atHome Team

Someone needs to rate the accuracy of “balance” in this case, because Leapforce seems to define it as “blend”. I get the sense that Google isn’t all that motivated to increase quality in this particular case.


It’s a good day when Google gets some comeuppance.


Raters gunna rate.


Ugh, so… I’m conflicted about the discussion around the gig economy in general.

On one hand, as has been reported, these sorts of mundane jobs with harsh requirements aren’t just limited to the gig economy. Lots of piecework or menial task work is like this, but less and less of it is done “in the west”. It feels a little first-world-y to focus on one and not the other.

Second, I have some personal experience with this sort of thing - my family, growing up, owned a fourth-generation midway. The workers (carnies!) there were paid $300/wk (1995 CDN dollars, $450 today’s dollars) for their work. If you count them working 11am-10pm 3 days a week (when we were open), that’s about $13 an hour. Add to that another 20 hours for setup/teardown, that’s about $8/hr. But they had to live with the midway as it travelled (or commute themselves from wherever), and had tasks to perform (cleaning/maintaining rides, etc) outside those tasks, and were basically expected to perform it “for free”. Realistically, from sunup until midnight six days a week, they were working.

By most standards, they worked for a few dollars an hour, and collected “pogey” (seasonal employment insurance) over the winter. but many of them had been doing it for a decade or more, as a lifestyle more than anything, and midways wouldn’t be able to survive without these lifers (would you visit a midway that cost 2x as much as today?)

I mean, I believe that folks should have an opportunity to have access to education and training, and meaningful employment. The fact that my province is trialing guaranteed minimum income programs, at the same time as our universal healthcare and free medication under 25 helps families with no other options than this cheap, underpaying work survive.

The work has to get done, so there’s essentially three options, IMHO: 1) supplement income for these folks, as Ontario is trying to do, 2) increase costs for the goods/services provided by these workers so they can be better paid, or 3) move these positions to regions where cost-of-living is low enough that the income provided by these jobs is acceptable.

I vastly prefer #1 over #2, because the first option places the burden on the wealthy who pay more taxes, versus option 2 which increases costs for all consumers, including these very people we’re discussing, who need lower costs for goods, not higher ones. #3 just kicks the can down the road as regions develop and standards of living increase.


Or maybe, Don’t Be So Obviously Evil?


Yeah, the whole ‘don’t be evil’ thing seems to have been quietly swept under the rug several years, and many billions of dollars, ago.


A twisted pro-corporate version of this is how a neoliberal UBI will likely function in the U.S.: enough to pay for the bare basics for the many people who will never have a shot at full-time work, but not enough that a lot of them won’t also be competing constantly to pick up piecework, seasonal work and temp shifts for paltry wages.


I think that’s how it’s likely to work here, too, in the sense that the idea is something along the lines of “If you don’t work, you’ll still have a home and your staples. If you aspire to more than that, find work.”

The reality here is very different, though, with strong labour laws and an $11.40/hr minimum wage.

There’s a whole other issue I didn’t even discuss - a lot of these jobs only exist because they’re currently cheaper than replacing them with robotics or spending more money on refining algorithms. When that’s no longer the case the new problem is going to be what to do when there are less jobs than people.


Nothing new about that problem…

It’s been the central social problem for every nation that has passed into/through the Industrial Revolution.


The success or failure of a society’s solutions to that problem largely determine the quality of life in that culture.

The stubbornness with which a culture clings to norms formed in an economy of scarcity, even as it transitions to an economy of abundance, will largely drive income inequality and overall misery.


Yes. In the U.S., any such UBI will be accompanied by the insistence of conservatives that any minimum wage legislation will disappear along with SSI, disability, and really all other social assistance programmes. That’s before you’ll get to universal single payer healthcare like Canada’s becoming permanently banished from American political discourse.

Liberals and progressives should not imagine for a moment that a UBI in America will necessarily lead to Fully Automated Luxury Communism. A lifestyle closer to what we’re now seeing in China, albeit with less social and economic mobility, is what most Americans will see after the conservatives are through tinkering with it.


Well, for anyone who wants to see all people thrive. For the people who think the proper/best order of society is a handful of people exploiting the many, then having a hungry, competitive workforce fighting over one-too-few chairs in the game is a wage and labor control mechanism. Of course, it has to be “tuned” so that people have enough money to buy the widgets coming out the other end… funny that, the fact that “society” and “industry” might not have the same interests. You could be fooled, listening to the right…


The problem is that this then leads to Wal-Martification, where the existence of social services pushes down the amount of wages that a company is willing to pay. While I do believe in a social safety net, I prefer option #2 to #1, because #1 means that the product itself is being subsidized. If the product is essential enough that it needs to be subsidized, it should be subsidized directly, rather than through lowered wages/social assistance. If, on the other hand, the product is not so essential that it should be subsidized, then workers’ wages should be allowed to rise to a living wage, and the cost of the good or service to rise as well. Otherwise, it’s just a subsidy on the product by another name.

This is why I prefer the idea of living-wage UBI to the piecemeal social assistance that currently exists; if you get enough money to live by whether or not you do some horrible job, then horrible jobs will have to pay enough that people will want to work there.


This sounds a lot like an AI algorithm implemented with humans. Weirdly meta if they’re being used to train AI and the AI is then in turn training them.


Don’t be evil, just be corporate.
Fool the world with your own importance.

1 Like

Well said :exclamation:


Why, thank you!


I did this job for a couple of months.

It wasn’t nearly as bad or as hard as the article makes out, but perhaps that’s because I was working on Japanese content. Presumably, the more applicants there are, the more competition for available work, and the stricter the evaluations. Anyway, I thought the pay was reasonable for the work involved, although it was pretty boring - I blitzed through the evaluations, alternating between evaluating and exercising in order to stay awake.

It was only ever an interim job for me though, and I wouldn’t want to do it long-term. Some of the testing was pretty arbitrary. For instance, I remember one task: A user searched for “J-pop boy bands”, and three videos were search results. Watching the videos they were all K-pop boy bands. I figured the results were “bad”, but not “worthless”, because there is crossover in popularity between J-pop and K-pop and boy bands are boy bands. So I gave each result 0.5~1 out of 5. The evaluation told me I’d made a serious mistake and should have rated each video as 0 out of 5. It’s weird trying to apply “objective” evaluation to seemingly subjective criteria.