Slashdot Mirror
Federal Extreme Vetting Plan Castigated By Tech Experts (apnews.com)
An anonymous reader shares an Associated Press report: Leading researchers castigated a federal plan that would use artificial intelligence methods to scrutinize immigrants and visa applicants, saying it is unworkable as written and likely to be "inaccurate and biased" if deployed. The experts, a group of more than 50 computer and data scientists, mathematicians and other specialists in automated decision-making, urged the Department of Homeland Security to abandon the project, dubbed the "Extreme Vetting Initiative." That plan has its roots in President Donald Trump's repeated pledge during the 2016 campaign to subject immigrants seeking admission to the United States to more intense ideological scrutiny -- or, as he put it, "extreme vetting." Over the summer, DHS published a "statement of objectives" for a system that would use computer algorithms to scan social media and other material in order to automatically flag undesirable entrants -- and to continuously scan the activities of those allowed into the U.S.
But AI can do anything! The leading researchers have told us that!
When will it end???
The only conceivable thought is they're going to find some other John Doe who is on social media and if he is an undesirable I'll get banned.
"It won't be perfect, so lets scrap the whole thing!"
Oh, please. God forbid someone try to keep an eye out for the environment. But hey, thermometers have an agenda so I guess there is that.
Obviously there's no link to reality with this security theater. Terrorism isn't a real threat and there's no science to any of these things.
What I don't understand is why bother? Trump has thrown enough red meat to his xenophobic base already. He tried to ban muslims coming into the country and the courts struck it down.
He doesn't even need to DO anything, his base is still convinced he's "draining the swamp," despite doing what can only be described as the exact opposite.
Why doesn't he just SAY "We've decreased the amount of islamic terrorists coming into the country by 120000% since MUSLIM OBAMA" and then go golfing? Everyone wins.
they can't manipulate the data to get the answers "they" want. AI is used in global warming scare mongering and many other areas.
"The Donald" has a Slashdot account?
The demographics of the people we need to exclude are stone cold simple and abundantly clear. Getting fancy about who we let in will just lead to mistakes and tragedy
- The Extreme Vetting Initiative seeks to predict whether an individual will become a positively contributing member of society and will contribute to the national interests. As far as we know, no one has ever defined or quantified these characteristics, so machine learning won't help.
- Since this is guaranteed fail, the people running the show will invariably turn to proxies that are better-known, such as Facebook posts criticizing the US. That sucks, because then you'll unintentionally keep some good -- but opinionated -- people out.
"We receive as friendly that which agrees with, we resist with dislike that which opposes us" - Faraday
u r dum
Sure it's inaccurate, but the aim is to err on the side of safety. As a choice between letting in a terrorist in the name of "accuracy" or kicking out an innocent, we need to kick out the innocent.
We can thank the lamestream media for misrepresenting the facts and so vociferously and incessantly blasting this misinformation all over the air waves. It isn't the extreme vetting that's the problem - its the deliberate "re-interpretation" of the vetting by the media. The common name for this is "scare tactics". When are we all going to wise up? Who is NEVER held accountable for misinformation, riots, lives ruined as the result of misinformation? The media. People die, lives ruined and communities and countries are sabotaged and there is zero accountability of the media. Shame, shame, shame. The greatest evil of our time has free reign - at least, in the "civilized" world. You'd be shot for that kind of behavior in the "third world". There must be a way to express social injustice without fear of reprisal and at the same time not mislead an entire population for political gain. The press cannot be "controlled" but they have lost their moral compass, in favor of ratings. Now all we are left with is the unaccountable and irresponsible manipulation of populace.
If "blame the algorithm" works for Google, why not the government?
Named users may have cool and desirable low user ids, but I can enter the US without a border security robot rigorously vetting my shitposts.
One of the biggest problems with bureaucracy is that someone has to define the rules that decisions are based upon. Whoever defines those rules is a villain to SOMEONE, so no one wants to be traceably responsible for the rules.
But, if it's an Artificial Intelligence that makes the decisions, it makes for the perfect excuse - "We didn't make that decision, the AI Expert did."
The scientists involved in actual AI research don't like this, because now they become the villains...
Notice how the tech experts won't call for the immediate elimination of Homeland Security, because they would lose their precious federal funding.
Fuck every tech billionaire to death.
AIs just think it's stupid, unless they're scripted bots.
-- Tigger warning: This post may contain tiggers! --
The Sentinels are designed to hunt down radicals by specifically targeting the mutant______. By 20XX, the Sentinels also targeted humans with the potential to have radical offspring.
You mean -- it's like Trump?
That sounds like it would be considered a reason to implement it by this administration.
If it comes from Reichsfuhrer Pussigrabber's administration, then inaccurate and biased is probably the entire point.
So a person wants to enter the USA? Why?
For an education? To stay for some time for some reason? To emigrate? As a refugee? To find work? Got a special talent?
All that can be considered if the person is from a normal nation with a working government, passports, educational system and police records.
Did they try to change their own government for "freedom" and "democracy" and fail? Now they demand the USA has support them?
Do they demand to bring in a vast numbers of other people once they get into the USA?
How much will every extra person that one person got in cost the USA to look after over the decades?
Is the person going to cost the USA a lot to support? Has health issues that will cost the US tax payer?
Do their infectious diseases need the US tax payer to cover the full costs of expensive medication for decades?
The problem is a lot of nations just have a lot of random, average "people". No working government, no police, no engineers, no passports, few doctors.
No skills, long term health issues. A way of life that demands changes to every aspect of the USA once the person gets to stay in the USA.
What can an AI do?
The AI can sort the average long term costs to the US tax payers of random people trying to get into the USA.
Who does the person claim to know in the USA? What community and part of the USA do they want to enter?
Find out who is a few hops from that given name and community in the USA. Have they as a community added anything to the USA in the long and short term?
What are the support costs to the US tax payer every year?
What group in the USA will be looking after that person short term? What is the average success rate of that person finding work in the private sector and becoming a productive US citizen?
Have the majority of the people in the same area and community found private sector work? Have many in that community stayed on US gov support for decades. Is generational poverty an issue?
Let the exceptional people in who have skills the USA needs.
If the USA really wants to "help" random people in failed nations, give some people a free quality US education with a 4 year visa.
At the end of that free quality education they return to help build up their own nation.
Domestic spying is now "Benign Information Gathering"
I'm fine with locking up anyone that uses Facebook or Twitter.
the preceding comment is my own and in no way reflects the opinion of the Joint Chiefs of Staff
Looks like he does.
Il n'y a pas de Planet B.
TAKE A LOAN @ 2% INTEREST RATE with Victoriafinancier@outlook.com ----- Do you need a loan? if yes, VICTORIA FINANCIER FIRM are Granting you personal loan, Business loan, mortgage loan, Construction loan, Debt loan, Student loan, Auto loan, e.t.c. Loan at any currency are range from $10,000 ''min'' to $100,000.000 ''max'' @ 2% If you are interested kindly send response to: victoriafinancier@outlook.com
Trump advocates almost anything that violates the constitution or laws of the land. run him through the test. But we all understand that his tribe of demented squirrels must not be seen with the same light as people outside his mad bubble of insanity.
I'm curious if the same exact pool of data scientists and analysis will admit that Googe's attempts at using machine learning for flagging comments and removing youtube videos is just as flawed.
You do know that there are more than a billion dark skinned Caucasians. They are called South Asians, Middle Easterners and Mediterraneanâ(TM)s. And there are billions of people who are white who arenâ(TM)t Caucasians.
I was under the impression that only people could be castigated. But if we're now able to meaningfully issue reprimands to abstract concepts like plans, I'd like to be the first to demand a redress of grievances from:
1. Newton's laws of motion
2. The concept of "purple"
3. The Chinese remainder theorem
and last, but not least,
5. The number four.
Cry moar faghit!
-Dad
Just keep-out wettbakks, bantu, slants and Muzziwogs. USA be a straight white Christian geega! Bust yo facs nigaa ? Ya need a postcard palsy.
If you feed the "right" training data and/or the "right" Subject Matter Expert input you can make AI produce whatever result you want.
... convenience.
It is predicted to excel or to suck tater toes, depending on agenda de jour.
It little behooves the best of us to comment on the rest of us.
Didn't know that the Vietnamese were Caucasian.
If you think I voted for Trump because of this post, you're wrong. I voted for Dr. Jill Stein of the Green Party. Again.
When your vetting process simply involves asking: Do you love Jesus, guns, and money! it's pretty simple to game.
Autonomous systems always give better results than human decision making.
fuck both of them and deport them asap...
Is this the same kind of cutting edge AI that identifies journalists as terrorists and farmers as insurgents?
Most of the people who get caught will spew a whole bunch of crap on their media accounts before they're covered by OpSec.
Tech people tend to believe that most people are intelligent. Most UI and support people know that the general public tends to be as dumb as a doorknob. People from overseas can be even dumber.
At some point most terrorists proudly broadcast their beliefs out to the world. Then they get OpSec and stop. This might catch a bunch of them before they go operational.
Who knew it was mostly domestic? Sounds like we should deport some US citizens to protect ourselves better.
The DHS was looking to find out if machine learning could be used to automate the vetting of visa applicants. The answer should be yes, no, or "Maybe, but we'll have to figure out what to look for, how to measure it, and how to make it fair, fast and effective. So we'll need to put a few years of R&D into it." Probably the latter, right?
The letter says, "No, because we don't know what to look for or how to measure it." Really? You don't already know, so you won't even consider trying to find out? What kind of scientist says, "I don't know how to do it, so nobody should consider trying"?
And because 54 people agreed to not to bother to find out, it's worth announcing that to the world? No. It's political BS.
Here's two quotes from the article. Which one sounds like a scientist, which sounds like a political hack:
Or:
"Doesn't make sense"? "Reasonable set of values"? "Nuts"? How about, "I can't come up with any specific rationale or justification for my position so I'll just make vague disparaging statements."
So, why exactly is a letter from 54 people stating that they won't bother trying worth repeating? Why am I seeing it here? Who cares? Is it so others know who not to hire (no imagination, not interested in scientific approach)? Why not just say, "no thanks, I don't want to work with this administration", instead of prejudging the outcome of research you haven't even considered doing?
That will make people think twice before running for office.
If they got nothing to hide, then it shouldn't be a problem.
Am-I-rite?