With this data, we were able to reconstruct Rotterdam’s welfare algorithm and see how it scores people. Doing so revealed that certain characteristics—being a parent, a woman, young, not fluent in Dutch, or struggling to find work—increase someone’s risk score. The algorithm classes single mothers like Imane as especially high risk. Experts who reviewed our findings expressed serious concerns that the system may have discriminated against people.
Are those patterns, indicators, of higher risk? That the results are then concentrated in certain demographics isn’t discrimination it’s reality. Or, if you prefer in economic language, the differences between taste discrimination and rational discrimination.
Do note, if AIs don’t reflect reality – however reality might have different outcomes for different demographics – then AIs are useless.
The problem with this sort of AI is that it’s backward-looking. For example you find that Moroccans in Rotterdam tend to be benefit cheats, so you look exclusively at them; and you miss the emerging trend of Somalians cheating on their benefits.
Humans are vulnerable to this type of mistake too; c.f. all government policy.
“then AIs are useless”
Well sorta. It’s more that they will quickly stop being used…
Perhaps I’m misinterpreting something, but do we really pay professionals to perform studies that tell us that “struggling to find work” leaves one at a higher risk of requiring welfare?
I got into the wrong field, obviously.
I suspect this indicates why AI will never be allowed to catch on: it keeps producing output based upon reality, and not upon woke doctrine.
See car insurance rates, acturarial tables, cervix smear tests for men, and a thousand other cases where reality won’t comply.
@ Tim the Coder
Your examples show that AI will be allowed to catch on but will simply be ignored by governments as actuarial tables are in setting gender-free annuity rates.
Before feminists insisted on “equality” and the government was forced to equalise state pension ages for men and women the state pension scheme was far *more* biassed in favour of women (it still is but less so).
This does not prove that women in general are less good at maths, just that the likes of Harriet Harman don’t care about maths.
“The algorithm classes single mothers like Imane as especially high risk. Experts who reviewed our findings expressed serious concerns that the system may have discriminated against people.”
If so, then the best way to prevent such discrimination is to have a system that doesn’t allow people such as Imane to live in Holland. What the flip is she doing there anyway?
‘the best way to prevent such discrimination is to have a system that doesn’t allow people such as Imane to live in Holland’
You’ve got the answer Sam!!
Hmmm.. Being in the same welfare system, getting the same benefits in the nextdoor region, just reading what they write about the cases has me going: Hmmmmm, yuuuuuup, no wonder you got flagged.
Bijstand is …in some ways pretty permissive.. but the Rules are strict. And you’d better Learn the Rules and Keep To Them if you don’t want to be put on the Grill.
And even then it’ll happen anyway, because Bijstand is not meant to be Easy Living.
You will be called in periodically, quite often if you just have been granted it, but you can expect to be checked once or twice a year, and be called in with a sheaf of paperwork every other year. Algorythm not needed.
Take “Imane”. The clincher there is the “cleaning jobs” she hops in and out of because of her “chronic back pain and dizzyness”.
You are only allowed extremely limited extra income in Bijstand, and you have to declare everything you earn that way. And underhand cash? ohboy…
With the amount of shenanigans in the cleaning industry that’s a Red Flag in itself. For blokes it’d be “odd jobs for a private contractor”.
You will be Grilled and X-rayed.
And yes, the boohoo about her having to log in to her bank account under scrutiny, true, and they have the right. It’s extreme, but mme. “Imane”s story tells us that she has been Naughty before.
They don’t hand out one month’s suspensions like candy, that’s two steps up from your First Warning ( reduction to 70% for a month.)
Pepita Ceelie, filling in the numbers.. Female, chronic back pain, exhaustion. The Holy Triad. Better have a medical attest that this is indeed the case.
Got picked for receiving a minor amount from her brother. Yes dear.. The Rules were No Donations From Family. In goods or cash. Period. That’s where the case of “having to pay back groceries” comes from.
Income from properly declared (volunteer) work? No problem up to [amount]. Family sending you even an incidental pick-me-up? Brrrrrt!
And yes, dear… To the System ( “AI” or Human) you are a Number. Part of the package, deal with it.
“The Netherlands takes a tough stance on welfare fraud, encouraged by populist right-wing politicians. ”
Errrmm… no.. The tough stance comes from huge amounts of fraud in the past. It was put in place well before the populist movement gained ground. With strong Socialist support, in fact.
The populists do keep the Modern Socialists from tweaking the system to benefit the “Focus Groups”-du-Jour they pretend to represent.
The populist parties haven’t even been in a position to do anything to the benefits system to begin with. Not enough seats in Parliament. Not enough seats in municipal councils.
Loud, yes, but representation-wise about as big, and loud, as the combined Socialists. Balance, you got to love it..
Oh, and the Rotterdam system didn’t get taken “offline” because it didn’t work as intended, as is ….suggested… in the article, conflated with the Child Support FUBAR by the Revenue Service.
It got taken out because due to WuFlu and new regulations the Rules changed. A lot.
And the system could not be retrained within the new rules and Not Bugger Up, so they decided to build up a new one, trained to the new Rules.
You really don’t think they’d reveal all the determinative factors and documentation for all the Smartasses to figure out if they were ever going to use the thing again, do you?
But all the Sad Cases in the article? Would have been called in even without any “AI” algorythm.
The only thing that thing does is make it a lot harder to get away with “bending the rules”. Which is a Good Thing.
“What the flip is she doing there anyway?”
Got imported with her parents, in the ummm… Second Wave Moroccan “temporary resident” contract labourers, judging from her age. Late ’70’s, early 80’s.
Who, of course, never left and brought Family… And never “integrated”.
I very carefully left stuff out from my post above, as that would be “racial profiling”. Cliché’s exist because, and all that.
That said.. She’s as “Dutch” as I am. The reason she’s here is because our collective (grand)parents failed to execute the “temporary” part of her parents’ contracts.
Because our politicians didn’t have the backbone then as now, and the roost was effectively ruled by the globalists and europhiles back then.
Morocco refusing to take them back, may have had a hand in that as well. Something only “resolved” ( let’s see if they keep their promise…) only mere months ago.
And that hasn’t even solved the problem of the dual passports, as Morocco has the US-ian attitude towards nationality, and denouncing your moroccon passport/nationality is a capital crime over there..
That particular part she can do nothing about even if she tried…
Thanks Grikath
Fascinating article. I don’t often click through.
I’m amused that these welfare peeps think that they are so charming that a real person would give them more money than a pocket calculator (algorithm). It would be easier, in fact, to get Somali welfare officers to interview Moroccans and vice versa. Giving the welfare officers a budget and a bonus for underspending.
But that would be racist. Score one for the machine.
Reality is profoundly racist, which is a problem for foolish white people desperately trying to pretend otherwise.