Given that almost the entire US population consists of immigrants or descendants of immigrants it should be a very simple algorithm: guilty. Just accept a relatively few false positives and the job's done.
US govt to use software to finger immigrants as potential crims? That's really dumb – boffins
A group of 54 computer scientists and academic researchers on Thursday asked the US Department of Homeland Security to rethink its plan for employing software algorithms to determine whether immigrants to the country should be admitted or deported. To implement various White House executive orders to limit immigration through …
COMMENTS
-
Thursday 16th November 2017 23:56 GMT Anonymous Coward
"The agency said it's looking for a system that allows it "to assess whether an applicant intends to commit criminal or terrorist acts after entering the United States.""
Translation: Are you or have you ever been a member of the Muslim faith?
Sure I have heard something like that before....
-
-
Friday 17th November 2017 00:49 GMT Kernel
Re: "generate a minimum of 10,000 investigative leads annually"
"So instead of actually looking for terrorists they're going to 'create' 10,000 of them each year ?"
Having spent the early part of my career in a public service role, I can tell you exactly what is happening here.
Someone has had a rush of blood to the head and has subsequently managed to create a cushy job, probably a couple of grades above where they were, to manage the new department of "Terrorism Pre-Convictions*".
In order to maintain the grade (and salary) in their new position they need to have a certain number of "investigators" working under them, which in turn means they need to have about 10,000 leads a year to investigate in order to justify the budget - but, and this is the important bit, the instigator of all this doesn't want 15,000 or more leads each year, as that would require significantly more staff and push the 'in-charge' role up another grade - and this would be a bad thing, because while you could swing a two grade jump in status, three grades is almost certainly going to attract the attention of someone who's already higher than you and you'll miss out.
Don't miss next season's exciting new top rated shows, 'T.P.C' and 'T.P.C Miami'.
-
Friday 17th November 2017 06:44 GMT Anonymous Coward
Re: "generate a minimum of 10,000 investigative leads annually"
... because while you could swing a two grade jump in status, three grades is almost certainly going to attract the attention of someone who's already higher than you and you'll miss out.
At least at the place I used to work, three grades in one year was simply impossible - the promotion rulebook rules did not allow for it. The circumstances simply did not matter - you could win a nobel prize, you could stop a minor war, or you could save your depatment billions. All you could possibly get is a double increment. Assuming that you have the paperwork to prove your teamwork and leadership ability - plus the rest of the checklist, of course.
-
-
-
-
Friday 17th November 2017 11:57 GMT DropBear
People's failure to learn is a built-in, inevitable direct consequence of human mortality and a relatively short lifespan. People live just long enough to be allowed meaningful discovery and gathering of experience that they then attempt to pass on before they die - but the next generation brings a fresh set of eyes and the "passing on" bit happens selectively and critically.
Often this is a very good thing, or else the immortal Ur-humans would have concluded long ago that the Earth is flat and the Sun is orbiting around it and that would be that, case closed. This way at least human knowledge gets a chance to progress every time the "old guard" finally kicks the bucket - and no sooner, as often observed with scientific dogma that stubbornly persists for a generation.
On the other hand, it is also often a very bad thing, as people who feel no need to actually fact-check things arbitrarily decide that inconvenient stuff they didn't personally see never happened, whether it's the moon landings or the Holocaust. And every time they think "surely WWII camps for Japanese couldn't have been that bad..." they get one step closer to willingly doing it all again...
-
-
Friday 17th November 2017 00:10 GMT Shadow Systems
An easy way to start...
Make it start by scanning every member of the current government & anyone associated with it such as military contractors. Check everyone whom it thinks is flagged as a bad person, then validate the results.
What's that, you can't scan folks you're beholden to for crafting the laws that don't apply to you but only the rest of us?
Go stick your head in a pig.
Or, you know, don't consider obviously brainless ideas as viable? Nahhh, that makes too much sense.
-
-
-
Monday 20th November 2017 19:10 GMT Anonymous Coward
Re: Perfectly logical
Sure, sure-- this algorithm is too non-specific to be exclude people reliably because it omits the "experimenter bigotry" variable. That's all politics and fake news, though.
We are much more interested in tremendous economic gains. That's why we've built a half-assed algorithm that may or may not facilitate the inclusion of people with high economic potential. If their weird Asian language or curry smell or funny music bothers some real Americans after immigration, well... there's no algorithm for that
-
-
-
Friday 17th November 2017 07:28 GMT MonkeyCee
Pre-cog
I do love the ideas that come out of government. In this case, please create a "pre-cog" machine that can accurately predict (for an individual) whether they are going to commit a crime* in the future.
And once we've built this amazingly powerful piece of technology, we'll only use it on..... the section of the population with the lowest criminal activity**. Not the politicians, not the judges, not the police, not the army, not the general population, just first generation immigrants, at the point of entry.
By the way, we already do have quite accurate predictions, based on populations rather than individuals. So if American weapons are supplied to oppressive regime A, a certain number of the people fighting A will add America to their list of countries to fuck with. However, pissing people off is considered an acceptable side effect of "redistributing the wealth" from brown people with mineral assets to white people with material assets.
* WTF is up with "crime and terrorism". Any act of terrorism is automatically a crime, albeit one with a political motive.
** 1st generation immigrants are much more law abiding than the general population. 2nd gen are about as law abiding as the general population.
-
Friday 17th November 2017 20:34 GMT Anonymous Coward
Re: Pre-cog
** 1st generation immigrants are much more law abiding than the general population. 2nd gen are about as law abiding as the general population.
I can't cite specific peer reviewed or official stats to to support this view, but observationally that conclusion is based upon certain conditions for migrants to migrate, and for immigration to be permitted. I would agree that it used to be true. But looking at the situation in the UK and Europe, the latest wave of migrants appear less likely to be law abiding than the general population.
-
Friday 17th November 2017 08:28 GMT MonkeyCee
But the current system is already excellent
Reminds me of the old story.
A kindly grandmother is visiting the USA for the first time. She gets to immigration, who ask her "Are you planning to overthrow the government of the USA by force or by subversion?" she thinks for a moment and replies "I do not condone violence, so by subversion then"
-
-
Friday 17th November 2017 09:27 GMT Aqua Marina
I've seen the algorithm
if skin = brown then Terrorist();
If religion = muslim then Terrorist();
if person = coming-from-far-east then Terrorist();
if person = american-with-personal-arsenal-and-psychotic-tendencies-occasionally-demonstrating-violence -and-might-just-shoot-up-a-school then Perfectly-Harmless-Person();
if skin = black then Shoot();
No honestly!
-
Friday 17th November 2017 10:03 GMT Anonymous Coward
Do it! Do it!
There is a definite upside to this. The people in ICE who are proposing this plan ought to be charged with implementing it. That is, they personally should be made to write the software and test it, until they have ironed out all the bugs and demonstrated that it works reliably as described.
Then (1) they will have learned a lot more about what computers can and can't do, which will make their future proposals more realistic; and (2) they can solve all the world's other hard computing problems with relative ease.
I especially enjoyed this bit:
'The researchers note that characteristics sought by the government – e.g. whether an individual will become a "positively contributing member of society" – are ill defined...'
It would be highly amusing to see the proposed software in action. What would be its verdict on the various members of Congress? On the Supreme Court? On American's many billionaires? On Mr Trump? On the DHS itself?
"He that is without sin, let him cast the first stone".
Ho ho ho.
-
Friday 17th November 2017 16:49 GMT Mike 16
Re: Do it! Do it!
@archtech:
There is a definite upside to this. The people in ICE who are proposing this plan ought to be charged with implementing it. That is, they personally should be made to write the software and test it, until they have ironed out all the bugs and demonstrated that it works reliably as described.
-----
You are assuming that none of them have a middle-school age niece or nephew to tell them about Stack Overflow, whereupon their work will be essentially what the hired amoral minions would have produced.
As for "testing"? with vague enough criteria anything will pass. Well , maybe some things will change. After all, a stopped clock is right twice a day, but a smartwatch with a dead battery...?
Keep in mind that e.g. an implementation of nanosleep() that _never_ returns complies with the spec. and the spec for this job will be a tad more complex.
-