> Even if surveillance is already widespread, that doesn't make more surveillance a good thing.
I'm one of those types who uses cash, doesn't own a cell-phone, doesn't have Facebook or any of the others, opts for pat-downs at the airport as a form of protest, etc. etc.
Having security and surveillance on our ports of entry seems entirely reasonable to me.
The problem I can imagine is that we're almost guaranteed to run into No-Fly-List style mistaken-identity situations and I sincerely doubt CBP will have reasonable recourse for such mistakes.
My privacy-related concern is that historically, surveillance data tends to be kept for long periods of time and used for purposes other than what it was collected for. My other concern is that I don't really welcome immigration officials having any more tools, giving how they're abusing the tools they already have.
That's a very fair and real concern, given "Synthesis Centers" and somewhat frequent allegations of parallel construction.
I suppose the proper answer is "Unfuck our security apparatus" but absent that, I see how abstaining from extremely abusable tools, e.g. this facial recognition, is a reasonable response.
I got tickets at the airport last week (cool little trick to save some money on taxes vs. online)...
The ticket lady told me some customer was upset with her. So she called over TSA security and was able to put him on the no-fly list. She just thought it was great! That blew me away. Someone can ruin a person's life in seconds. Punchline: be a good, politically correct person at the airport. Even the hint of anger could get one on the no-fly list.
> Having security and surveillance on our ports of entry seems entirely reasonable to me.
Having "ports of entry" as a concept seems entirely unreasonable to me.
I can travel between Bristol and Manchester without having to justify myself to anyone, prove who I am, etc. and I don't see why I shouldn't be able to travel between England and Canada under the same circumstances. Everyone in the world is an equally valid human being. How can one human being say to another "you're not allowed to exist in this public place unless you've got the correct paperwork"? It's barbaric.
> CBP is testing biometric scanning at a dozen or so U.S. international airports to ensure that people leaving the country are who they say they are, and to prevent visa overstays. The Transportation Security Administration, another agency within the Department of Homeland Security, is testing similar devices at security check-in lines ... CBP hopes to have face scanners installed at all the nation's airports in four years.
> Even if we include costs of testing, figuring out AWS and actually running the facial recognition on our scenario, it’s going to be under $10 ... the ACLU found the Orlando Police Department spent just $30.99 to process 30,989 images .. Compared to other facial recognition projects currently being run by the federal government, the Amazon service is staggeringly cheap.
It is a little, unsettling, when they set up checkpoints to make sure people exiting the country are authorized to do so. I'm sure people can state good reasons, but it feels very East Germany-est.
At first this didn't add up, but I think they want to be able to prove that you overstayed your visa. But with the current process, you can come, overstay your tourist visa, and leave, but they don't stamp you on departure, only on arrival, so they can't actually tell that you left.
Again, is this a real, rampant, dangerous problem? Probably not. But it's a real use-case (so they don't have to give fake reasons) for state-wide surveillance. Since it also claims to target foreigners (which by definition aren't voters), most locals/voters/politicians won't care, thinking it doesn't apply to them.
Not that I'm defending this, I think this level of surveillance is extremely dangerous. And the most obvious thing is that they'll slowly and quietly expand this to everyone else.
It has kind of been a thorn in the side of immigration enforcement though that determining who has potentially overstayed their visa because of the lack of exit tracking. Because of that trying to enforce visa overstays was really hard because you just have a huge list of people who have come with not much info about who actually left so trying to check up on visa overstays would waste a lot of time on people who had just come and gone as required. This trims down that list to be muuuch smaller so they can more easily check without wasting a lot of time.
Not long ago even European countries had double borders - you had to go through one checkpoint of your own country, that would verify that you can leave(you are not a wanted criminal for example), and then another checkpoint of the country you wanted to enter, to check if you have the right to enter it. The area between the two was the no-mans-land. This arrangement still exists between some South American and African countries, you have to go through two separate checkpoints.
> Not long ago even European countries had double borders
Our current situation might be marginally better than "double borders", but you're probably aware we're headed in a police statey direction all over the world.
It's not to make sure they are authorized to leave. It's for matching exits to earlier entrances, so they can tell if people who came on time-limited visas left when they were supposed to.
The US and Canada haven't had exit checks in the recent past. What they've done is collect flight manifests and use that as confirmation that a person has left the country.
However, it's quite common in Europe and other countries. Every time I fly out of a European county, my passport has been check to confirm that I actually left. Same with several Asian countries.
Apparently the UK doesn't do it, but other EU countries do.[1]
> Apparently the UK doesn't do it, but other EU countries do.
The UK has the ID document details checked by airline staff (or their contractors, like handling agents) with document details passed on to the Border Agency.
We need more details. It doesn’t state if they would do anything at all. They could be tagging faces for visa overstays so they would be flagged next time they can through.
It sounds like they don’t really have any automation to relate departures with overstayed visas.
The more hostility the public has towards immigrants, the more systems are set up to track people and deport them at the first opportunity, the first error that causes someone to move from "legal" to "illegal".
These tracking systems end up looking at everyone, because anyone might be an immigrant.
It’s one thing to obscure an object designed to be recognizable, and another to regulate clothing, makeup, etc. if it’s not a mask, it’s going to probably be legal. The first amendment is very broad, and what you wear has been protected many times under it.
It’s not that simple. If the police or border patrol want to ID you they can compel you to remove a mask. There are also various regulations around wearing full face masks in banks and the like.
I believe it's very simple. If a police officer wants to do just about anything with you, they can and will, regardless of the rights you have, or restrictions put on the police.
Also, the Bill of Rights trumps regulations, but a bank is a private entity and you don't have rights there.
Modern facial recognition systems are actually pretty accurate. The problem is when you use them for screening on a population where only a small fraction of the people fall into the group you are screening for. You can then run into the "False Positive Paradox".
This is more well known in the context of medical testing. Suppose you are screening for a rare disease. You are testing a group of 10 000 people, of which 100 have the disease and 9 900 do not.
Let's say you have a test that is right 99% of the time.
Of the 100 people you test who actually have the disease, your test will catch 99 of them.
Of the 9 900 people you test who do not have it, your test will falsely report that 99 of them have the disease.
The result is you get 198 people reported as having the disease, but only 50% of them actually have the disease. The other 50% are false positives.
If the test were right 90% of the time rather than 99% of the time, the results would be 90 positives who have the disease, and 990 positives who do not. Only 8.3% of the people who test positive actually have the disease!
If you used that 90% test on a population where 5 000 of 10 000 have the disease, things are dramatically different. 90% of the people who test positive actually have the disease.
Same thing with facial recognition. Set up a facial recognition system at a big football game and look for people who have outstanding arrest warrants, and you are in a situation like the medical test for the rare disease. A 90% accurate system might give you 10 times as many false positives as true positives.
Set up a system at a border that has a lot of one-day tourist traffic both ways, and use it to try to match up people coming back at the end of the day with people who left earlier in the day, and it might be more like the 50/50 disease. A 90% accurate test might only have a 10% false positive rate.
I'm not sure what the general objection is to facial recognition technology. What's the difference between this and hiring a team of autistic savants who can remember a ton of faces? They use their eyes (cameras) to look at faces, and their brains (algorithms) to classify the image. One just costs a lot less to create.
‘What is wrong with a surveillance balloon hovering over every city following all citizens and logging their activity? It’s no different to the police officer on the corner observing the public’.
It’s a matter of scale, efficiency and generality.
The scariest thing imaginable is a truly efficient government. Our system works because there is room for rebels and alternate ways of thinking and living to bubble into the whole polity.
Gay rights for example.
If the government had had the ability to monitor all homosexuals back when the laws were truly terrible - they’d all be in prison.
I think it's perfectly correct to believe that the US government in its current state is pure shit for more than just the reasons you mentioned. But forced inefficiency sounds like a ridiculous way to go about fixing it.
If society takes issue with the laws, fix the laws. But having a law that relies on selective and inefficient enforcement is a stupid law.
Searching and following a particular person with probably cause is fine under the constitution allowing the police effectively follow anyone 24/7 strips away a lot of 4th amendment protections. The fact that the laws haven't quite caught up to the ability of police to spy on people en mass should not be a justification for them doing so. Law makers lag behind the public and the state of technology by decades.
> The scariest thing imaginable is a truly efficient government.
but, doesn't it depend upon what the government's doing?
if the government is efficiently preventing mining companies from pouring cyanide into waterways, regulating large Wall Street banks and building safe and ubiquitous mass transit to replace large swarms of polluting vehicles -- is that scarier than every other scary thing?
You can't give it efficiency in some areas and hope it stays inefficient in others. I would rather have a government that is incapable for preventing the next major civil rights battle and let a bad corporate actor emerge every now and then, than have a government that is able to crush even the slightest populist uprising without missing a beat.
Because scale matters. Things that are acceptable when they're difficult and expensive become unacceptable when the cost drops towards zero.
When you can do things broadly and cheaply, you tend to do so just because you can. This is analogous to "data mining" in "big data", aka one major issue with modern science -- when you look at everything trying to identify basic patterns, you'll have a heck of a lot of false positives.
I'm pretty sure the CBP would give you the finger and ask you to "come and get it". I really don't understand why Europeans think other nations will follow other nations laws
I think some of the general weirdness people feel around AI policing is that there are so many dumb laws on the books, and if we let robots judge, then there is no mercy. Also, everyone will be in jail because, on average, every American commits a felony every day ( because there are so many weird dumb laws ).
I think someone should make AI to refactor law code into 1/1000 its current size. There are like millions of lines of law, even for stuff like soybean farming. I'm sure a lot of that is duplicitous, redundant legal speak, not a precise computer-type language.
Spoken like someone who's never seen a million lines of business logic in actual code.
And it develops for the same reasons: in a complex system there's going to be all sorts of weird corner cases, and some of them are bad, so you write some more code/laws to deal with those outcomes ...
How does that detract from the point? They are arguing for a sort of "code deletion" and "refactoring" of laws. A far superior way of maintaining a codebase.
Not just "ADD MORE LINES OF CODE", which you make sound is like some concrete reality of how things always get done. It's not. And no one should strive to do things this way.
Adding more lines of code is something that you can control and do right now. Changing the world so you do not have to do that takes significantly more time and effort.
In other words, growing complexity of the system to deal with new cases is past of least resistance.
If we always followed that logic we'd still be using first gen OS from the 1970's, or perhaps we'd still be using the Code of Hammurabi. Reevaluating legacy systems in the face of paradigm shift is a crucial part of natural evolution.
Except real life isn't as simple as even the largest repos of code. There are edge cases that deal with morality and ethics rather than if and for statements.
There is a certain amount of business logic that can't be refactored away ( -someone's law ), but there is also bloat.
Who is refactoring this law code? No one. It gets written once and stands forever. It gets repealed once and a while, but it doesn't get actively refactored.
Spoken like someone who's never experienced the inanity of the US's criminal justice system.
Many of the laws on the books are there due to lobbying, not logic, reason or need. Some are based on moral outrage to propaganda, some are due to businesses wanting greater leverage of their market and greater control over the entry-barrier. Many laws are totally unconstitutional but remain because someone must be successfully challenged which is hard when the law is meant to target the homeless guy you just don't like sitting on your corner who wouldn't have the means to hire a proper attorney.
I think what the original poster was getting at is that AI enforcement, with a broken legal system is a recipe for abuse and injustice.
The entire idea of selective enforceability is an affront to the concept of justice.
I've had my car searched for drugs by police in Arizona without cause other than they thought I "appeared to be on drugs" due to the fact they made me stand outside in the snow for 10 minutes while questioning me during a bogus traffic stop and I eventually started to shiver. I was driving through AZ on my way from Colorado to California... two places recreational marijuana are perfectly legal (by state standards) yet if they would have found even the smallest amount of weed in my car even if someone else had just dropped it somewhere (they check literally everywhere -- including rifling through my tea bags with their dirty paws), I would have been in jail on felony charges.
Granted I made sure my car was immaculate because I know how insane the people of AZ are when it comes to "criminal justice" and all it cost me was an hour of travel time. But the fact remains that there is no presumed consistency in law and felonies are a big deal and have consequences that reach far beyond some localized legal affair.
I have known people who have picked up felonies because they were doing something innocuous beside being in the wrong place at the wrong time. And when we start talking about AI enforcement, it's a scary concept because a felony means that for instance you could be picked up on federal firearm charges because some program found a slingshot in a picture of your bedroom on Facebook which gave them cause to issue a warrant to raid your house. And then you aren't able to flee the country because you can't just walk down to Mexico like you used to be able to because of the AI surveillance.
Taking the high-horse and saying people should trust in the system and just do their duty as citizens is great and all, but that type of ignorance is also why we have the highest prison population in the world. People who could be benefiting society even if only marginally, who are forced to be strictly welfare cases.
Not to mention, this isn't solving any problem - just opening the door to a bunch of new potential problems, so what's the point, other than lining the pocket of some random government contractor who probably lobbied for the tech in the first place? This is inane.
Big Company X: Speaks with big fancy words about technology that the government doesn't relatively understand, so they use that to manipulate said government into accepting the technology, because it sounds "fancy" and thus it must be good. Of course it would be spoken in a formal and possibly legal way too, which will cause even more confusion for the government department that the idea is being pitched to.
The result?
Expectation: A government that wants to adapt good technology and morally acceptable ways to have a great sociey.
Reality: A big company that "tricked" the government into thinking they will reach their expectation, when really they're just attempting to drain the government for its money, as well establish somekind of "power" in the society.
That's why we have judges (and why we should preserve the judiciary rather than undermining it with mandatory sentencing laws): to deal with the corner cases. The law could be kept simple and judges be trusted to deal with the complex corner cases.
Having said that, I guess all that happens is statutory law remains simple, but case law grows more complex as the body of judicial decisions grows? What we seem to be missing is a "refactoring" process to periodically rewrite statutory law to replace bloat in case law with a more otimised statute. Like coders, politicians prefer to work on shiny new features than maintaining existing features?
Are you really arguing that the law should be replaced with a precise computer language?
Oh my god HN... Laws are typically written with a lot of vague language because humans operate in vague, gray areas. If you write an entirely prescriptive set of laws (Like Hammurabi's Code of Laws) you will forget some things, be too draconian in others, and overall a sense of justice will be lost.
Not so much that.. but I bet there are logical contradictions in law, and repetition. You know how typescript can check for errors? I bet there are lots of errors in law code.
Lets say the same law says something is illegal and not doing something is illegal, because of some convoluted language and whoever passed the law missed it. Maybe not refactor right away, but at least check for errors.
Also, if a computer could process law code, lawyers would be obsolete, patent trolls wouldn't have leverage, etc etc.
It's interesting because your view of the law is very optimistic, and you assume law is deterministic.
The truth is that fundamentally law is not deterministic. You can't deterministically apply the reasonable person standard, or even the layman standard. How do you handle the expanse of human nature, and it's intricacies and vagaries?
Law can't be perfectly deterministic because life is not deterministic given the constraint of free will. Trying to make an idealized code without human judgement (and yes, that exposes human bias as well) results in the worst kind of law - an autocracy.
Law is logic. Logic can be refactored. I'm betting in 5 years we'll see lawyers use AI to help them do their job, specifically to help them create logical arguments for their cases.
What I'm suggesting is computers to take the logic of law code to poke around, tracing logic around and finding the obvious logical inconsistencies. A goal would be reducing law code as well as making it more laymen readable.
Obviously a huge benefit of being judged by humans and peers is a sense of mercy. I'm not saying we should mess with that. Also, I would say that anything the computer found would be a suggestion to congress as a way to refactor law, not trusting the computer to automatically do it.
Law is not logic in the way a computer scientist understands it. It's intentionally changed with tortured reasoning to get the outcomes the Supreme Court wants and seems fair. This is how the constitution can simultaneously allow segregation and 100 years later prevent it without a word being changed. The only change was the society living under it.
Law is logic but it's soft logic. So yes all we need is to get AI to understand the concept of mercy, basically to capture the spirit, the 'why' of the law instead of the 'what'.
A lot of the dumb laws where mercy is frequently allowed are that way because they were designed to be, as catch-alls for harder to define grievances. An AI 'judge' just needs to understand what it means to be a good person.
Of course then that could be expanded to an AI which simply tallies moral points and for which a given person committing a given crime gets a punishment designed to best rehabilitate them not to commit the crime again. But that would also be a system which is judge jury and executioner and not even human, and has some self-designed ultimate moral code...thats a lot to get right, and a lot to go wrong.
Law is fuzzy logic that takes in wide band information in a contextual way. AIs are terrible at that, at least as we have them now.
> I'm betting in 5 years we'll see lawyers use AI to help them do their job, specifically to help them create logical arguments for their cases.
So, actually, Lawyers are using AI right now - but not in that way. Specifically AI is used to analyze and bring up relevant case law, in the same way a law clerk might. It's not perfect, it still requires humans to clean up, but it saves some hours since it can analyze a lot of law quickly and vaguely run a likelihood analysis.
Law is also in English. English is not formally logical. Frequently laws have subtle wording problems.
Laws and punishments were created and tuned expecting that only a small portion of offenses and offenders would be caught, or that punishment could be applied selectively. Punishment of some plays a role as a publicly visible deterrent. That means the punishments are more severe then could be reasonably applied for every infraction of the crime. Crimes that are very common such as speeding or cannabis make good examples of this. If detection and prosecution of those crimes became 100% effective and consistent, courts would be overwhelmed, and the economy would be turned on its head as people go bankrupt paying fines and legal fees or losing their jobs.
I assume people would very move quickly to get the speed limits changed to a number that represents the actual maximum speed you can go before receiving a ticket, rather than the vague percentage/luck combo that it is now.
I'm not disagreeing with you, because there is definitely a balance, but I definitely see places where we have have too many laws, which leaves room for tyranny.
For example, If there are so many laws that everyone is always doing something illegal all the time, then a corrupt cop can stop anyone he wants to at any time, because they've got to be guilty of something. It's a reason for types of deregulation, and keeping law slim in general.
I think we're in agreement. Criminalizing common activity gives police discretion for selective enforcement, which leaves the door open for abuse. Contrived example: everyone is speeding, and a corrupt cop only tickets cars with bumper stickers for sports teams that oppose his favorite. It could be a lot worse than that, of course.
When police have to overlook a large amount of activity that violates the law, something is wrong with the law.
Have you read the laws on soybean farming or are you just making assumptions? Computer programs are small because they have limited forms of input. Because the input to law is the full universe of possible human behavior, law as code would be much larger than it is today, and would require pre-imagining every possible human action, rather than using a phrase like “malicious intent” and letting a judge and jury figure it out.
Your suggestion is farcical, both in the suggested method and size metric. Short laws would only lead to a greater degree of interpretation from fallible humans to reach a disposition.
The idea of "to add 1 law, remove 2 others" worked well in other locales. The problem of laws being so freeform (sometimes being broad or multi-laws in one) is part of the problem. Forcing succinctness is just stage one of cleaning it up.
> I think someone should make AI to refactor law code into 1/1000 its current size
That someone should "make AI to do X" certainly isn't too much to ask.
Likewise, why can't just someone make a pill that gives us perfect memory and understanding of all laws on the books? Make the pill so you only have to swallow it once, and then your knowledge gets automatically updated. Make a pill that does all that, but is also very easy to make and has no side-effects, so there are no big costs or complications. Such a complex problem, and yet we both proposed very elegant solutions within minutes of thinking about it. Why isn't this this sorted already?
You should reconsider your mental model of the law. With respect, it’s misinformed.
Law in its writing and interpretation is less deterministic, and quantity of statute is far less of a factor in its complexity than most in STEM realise.
Further, to parse law in a sense that it could be sensibly refactored, would be impossible with AI/ML as the field stands today, and I think will never be possible.
The way I see it is that your face is kind of like your license plate number. The government being able to map your likeness to your official metadata (taxes, license, etc) is somewhat legitimate. It's when they start tracking it everywhere that I have huge issues. Feeding security camera footage into facial recognition software is like putting up an ALPR at every street corner. It's not an invasion of privacy per-se but because technology allows all that metadata to be sifted through on demand at scale it's basically government stalking.
The facial recognition ship has long sailed for 1st world citizens thanks to how many pictures most of us have online, official photos for ID, etc, etc. The government can already map your face to your identity.
This particular program is probably more targeted at the non US residents crossing the border because the three letter agencies don't have access to as much good data and metadata on them. Because all the Mexican government agencies don't share ID photos with the three letter agencies the three letter agencies have to do it themselves hence facial recognition at borders.
As others have said I'd be very surprised if they weren't just trying to legitimize something they were already doing.
Exactly. Facial recognition already happens, but it's it's "natural", analog form. The border guard will make sure your documents are valid and the face in the document matches that of your face. That's more than adequate. If it's important to visa conditions, the date of entry/exit can be recorded, or for the sake of security, the date can be recorded in an offline log to be opened only by due process.
Photographing and storing data on millions of people without probably cause, sharing them with third parties (like Amazon) and keeping them forever, that's not just unnecessary, but also dangerous.
There was a prior HN discussion about a previously convicted felon and four-time deportee who still managed to re-enter via the regular border crossing and steal a million dollars in gold. This technology will solve that edge case.
If they're anything like the automated data collection kiosks at LAX, this is going to be a colossal failure. In my experience, they are terrible at identifying brown people and children. Good luck implementing this at scale.
101 comments
[ 4.7 ms ] story [ 170 ms ] threadCitation?
I'm one of those types who uses cash, doesn't own a cell-phone, doesn't have Facebook or any of the others, opts for pat-downs at the airport as a form of protest, etc. etc.
Having security and surveillance on our ports of entry seems entirely reasonable to me.
The problem I can imagine is that we're almost guaranteed to run into No-Fly-List style mistaken-identity situations and I sincerely doubt CBP will have reasonable recourse for such mistakes.
I suppose the proper answer is "Unfuck our security apparatus" but absent that, I see how abstaining from extremely abusable tools, e.g. this facial recognition, is a reasonable response.
I got tickets at the airport last week (cool little trick to save some money on taxes vs. online)...
The ticket lady told me some customer was upset with her. So she called over TSA security and was able to put him on the no-fly list. She just thought it was great! That blew me away. Someone can ruin a person's life in seconds. Punchline: be a good, politically correct person at the airport. Even the hint of anger could get one on the no-fly list.
Having "ports of entry" as a concept seems entirely unreasonable to me.
I can travel between Bristol and Manchester without having to justify myself to anyone, prove who I am, etc. and I don't see why I shouldn't be able to travel between England and Canada under the same circumstances. Everyone in the world is an equally valid human being. How can one human being say to another "you're not allowed to exist in this public place unless you've got the correct paperwork"? It's barbaric.
> CBP is testing biometric scanning at a dozen or so U.S. international airports to ensure that people leaving the country are who they say they are, and to prevent visa overstays. The Transportation Security Administration, another agency within the Department of Homeland Security, is testing similar devices at security check-in lines ... CBP hopes to have face scanners installed at all the nation's airports in four years.
AWS facial recognition service, https://www.forbes.com/sites/thomasbrewster/2018/06/06/amazo...
> Even if we include costs of testing, figuring out AWS and actually running the facial recognition on our scenario, it’s going to be under $10 ... the ACLU found the Orlando Police Department spent just $30.99 to process 30,989 images .. Compared to other facial recognition projects currently being run by the federal government, the Amazon service is staggeringly cheap.
Again, is this a real, rampant, dangerous problem? Probably not. But it's a real use-case (so they don't have to give fake reasons) for state-wide surveillance. Since it also claims to target foreigners (which by definition aren't voters), most locals/voters/politicians won't care, thinking it doesn't apply to them.
Not that I'm defending this, I think this level of surveillance is extremely dangerous. And the most obvious thing is that they'll slowly and quietly expand this to everyone else.
It's also a "permission to enter" some other country. But without it, you won't be flying anywhere.
Then you've got visas, which are "permissions to exist" in a certain geographical area (within a certain time-frame).
We've all had a default assumption that we can get a new passport any time, but it's actually never been guaranteed.
Our current situation might be marginally better than "double borders", but you're probably aware we're headed in a police statey direction all over the world.
However, it's quite common in Europe and other countries. Every time I fly out of a European county, my passport has been check to confirm that I actually left. Same with several Asian countries.
Apparently the UK doesn't do it, but other EU countries do.[1]
[1]https://www.flyertalk.com/forum/checkpoints-borders-policy-d...
The UK has the ID document details checked by airline staff (or their contractors, like handling agents) with document details passed on to the Border Agency.
It sounds like they don’t really have any automation to relate departures with overstayed visas.
These tracking systems end up looking at everyone, because anyone might be an immigrant.
Anyways, time to paint CV confusing patterns on cars, clothes and window glasses..
https://qz.com/878820/new-camouflage-promises-to-make-you-un...
We've got to somehow make this fashionable. Confuse the heck out of image recognition systems. At least the optical ones.
I believe masks are also legal. First amendment + a religion that lets you cover your face = your right.
Also, the Bill of Rights trumps regulations, but a bank is a private entity and you don't have rights there.
https://en.wikipedia.org/wiki/Anti-mask_laws
https://en.wikipedia.org/wiki/Development_mule
This is more well known in the context of medical testing. Suppose you are screening for a rare disease. You are testing a group of 10 000 people, of which 100 have the disease and 9 900 do not.
Let's say you have a test that is right 99% of the time.
Of the 100 people you test who actually have the disease, your test will catch 99 of them.
Of the 9 900 people you test who do not have it, your test will falsely report that 99 of them have the disease.
The result is you get 198 people reported as having the disease, but only 50% of them actually have the disease. The other 50% are false positives.
If the test were right 90% of the time rather than 99% of the time, the results would be 90 positives who have the disease, and 990 positives who do not. Only 8.3% of the people who test positive actually have the disease!
If you used that 90% test on a population where 5 000 of 10 000 have the disease, things are dramatically different. 90% of the people who test positive actually have the disease.
Same thing with facial recognition. Set up a facial recognition system at a big football game and look for people who have outstanding arrest warrants, and you are in a situation like the medical test for the rare disease. A 90% accurate system might give you 10 times as many false positives as true positives.
Set up a system at a border that has a lot of one-day tourist traffic both ways, and use it to try to match up people coming back at the end of the day with people who left earlier in the day, and it might be more like the 50/50 disease. A 90% accurate test might only have a 10% false positive rate.
You can reduce almost anything that way.
‘What is wrong with a surveillance balloon hovering over every city following all citizens and logging their activity? It’s no different to the police officer on the corner observing the public’.
It’s a matter of scale, efficiency and generality.
The scariest thing imaginable is a truly efficient government. Our system works because there is room for rebels and alternate ways of thinking and living to bubble into the whole polity.
Gay rights for example.
If the government had had the ability to monitor all homosexuals back when the laws were truly terrible - they’d all be in prison.
Runaway slaves? Same deal.
Murder and rape? Same deal.
I think it's perfectly correct to believe that the US government in its current state is pure shit for more than just the reasons you mentioned. But forced inefficiency sounds like a ridiculous way to go about fixing it.
If society takes issue with the laws, fix the laws. But having a law that relies on selective and inefficient enforcement is a stupid law.
Imagine trying to argue your point to the USA Founders, who seceded from the government they found oppressive.
but, doesn't it depend upon what the government's doing?
if the government is efficiently preventing mining companies from pouring cyanide into waterways, regulating large Wall Street banks and building safe and ubiquitous mass transit to replace large swarms of polluting vehicles -- is that scarier than every other scary thing?
When you can do things broadly and cheaply, you tend to do so just because you can. This is analogous to "data mining" in "big data", aka one major issue with modern science -- when you look at everything trying to identify basic patterns, you'll have a heck of a lot of false positives.
Edit: I previously wrote the CBP, but I meant the company that owns the facial recognition tech and stores my facial data.
I think someone should make AI to refactor law code into 1/1000 its current size. There are like millions of lines of law, even for stuff like soybean farming. I'm sure a lot of that is duplicitous, redundant legal speak, not a precise computer-type language.
And it develops for the same reasons: in a complex system there's going to be all sorts of weird corner cases, and some of them are bad, so you write some more code/laws to deal with those outcomes ...
Not just "ADD MORE LINES OF CODE", which you make sound is like some concrete reality of how things always get done. It's not. And no one should strive to do things this way.
In other words, growing complexity of the system to deal with new cases is past of least resistance.
Who is refactoring this law code? No one. It gets written once and stands forever. It gets repealed once and a while, but it doesn't get actively refactored.
Many of the laws on the books are there due to lobbying, not logic, reason or need. Some are based on moral outrage to propaganda, some are due to businesses wanting greater leverage of their market and greater control over the entry-barrier. Many laws are totally unconstitutional but remain because someone must be successfully challenged which is hard when the law is meant to target the homeless guy you just don't like sitting on your corner who wouldn't have the means to hire a proper attorney.
I think what the original poster was getting at is that AI enforcement, with a broken legal system is a recipe for abuse and injustice.
The entire idea of selective enforceability is an affront to the concept of justice.
I've had my car searched for drugs by police in Arizona without cause other than they thought I "appeared to be on drugs" due to the fact they made me stand outside in the snow for 10 minutes while questioning me during a bogus traffic stop and I eventually started to shiver. I was driving through AZ on my way from Colorado to California... two places recreational marijuana are perfectly legal (by state standards) yet if they would have found even the smallest amount of weed in my car even if someone else had just dropped it somewhere (they check literally everywhere -- including rifling through my tea bags with their dirty paws), I would have been in jail on felony charges.
Granted I made sure my car was immaculate because I know how insane the people of AZ are when it comes to "criminal justice" and all it cost me was an hour of travel time. But the fact remains that there is no presumed consistency in law and felonies are a big deal and have consequences that reach far beyond some localized legal affair.
I have known people who have picked up felonies because they were doing something innocuous beside being in the wrong place at the wrong time. And when we start talking about AI enforcement, it's a scary concept because a felony means that for instance you could be picked up on federal firearm charges because some program found a slingshot in a picture of your bedroom on Facebook which gave them cause to issue a warrant to raid your house. And then you aren't able to flee the country because you can't just walk down to Mexico like you used to be able to because of the AI surveillance.
Taking the high-horse and saying people should trust in the system and just do their duty as citizens is great and all, but that type of ignorance is also why we have the highest prison population in the world. People who could be benefiting society even if only marginally, who are forced to be strictly welfare cases.
Not to mention, this isn't solving any problem - just opening the door to a bunch of new potential problems, so what's the point, other than lining the pocket of some random government contractor who probably lobbied for the tech in the first place? This is inane.
The result?
Expectation: A government that wants to adapt good technology and morally acceptable ways to have a great sociey.
Reality: A big company that "tricked" the government into thinking they will reach their expectation, when really they're just attempting to drain the government for its money, as well establish somekind of "power" in the society.
Having said that, I guess all that happens is statutory law remains simple, but case law grows more complex as the body of judicial decisions grows? What we seem to be missing is a "refactoring" process to periodically rewrite statutory law to replace bloat in case law with a more otimised statute. Like coders, politicians prefer to work on shiny new features than maintaining existing features?
Oh my god HN... Laws are typically written with a lot of vague language because humans operate in vague, gray areas. If you write an entirely prescriptive set of laws (Like Hammurabi's Code of Laws) you will forget some things, be too draconian in others, and overall a sense of justice will be lost.
Lets say the same law says something is illegal and not doing something is illegal, because of some convoluted language and whoever passed the law missed it. Maybe not refactor right away, but at least check for errors.
Also, if a computer could process law code, lawyers would be obsolete, patent trolls wouldn't have leverage, etc etc.
The truth is that fundamentally law is not deterministic. You can't deterministically apply the reasonable person standard, or even the layman standard. How do you handle the expanse of human nature, and it's intricacies and vagaries?
Law can't be perfectly deterministic because life is not deterministic given the constraint of free will. Trying to make an idealized code without human judgement (and yes, that exposes human bias as well) results in the worst kind of law - an autocracy.
What I'm suggesting is computers to take the logic of law code to poke around, tracing logic around and finding the obvious logical inconsistencies. A goal would be reducing law code as well as making it more laymen readable.
Obviously a huge benefit of being judged by humans and peers is a sense of mercy. I'm not saying we should mess with that. Also, I would say that anything the computer found would be a suggestion to congress as a way to refactor law, not trusting the computer to automatically do it.
A lot of the dumb laws where mercy is frequently allowed are that way because they were designed to be, as catch-alls for harder to define grievances. An AI 'judge' just needs to understand what it means to be a good person.
Of course then that could be expanded to an AI which simply tallies moral points and for which a given person committing a given crime gets a punishment designed to best rehabilitate them not to commit the crime again. But that would also be a system which is judge jury and executioner and not even human, and has some self-designed ultimate moral code...thats a lot to get right, and a lot to go wrong.
> I'm betting in 5 years we'll see lawyers use AI to help them do their job, specifically to help them create logical arguments for their cases.
So, actually, Lawyers are using AI right now - but not in that way. Specifically AI is used to analyze and bring up relevant case law, in the same way a law clerk might. It's not perfect, it still requires humans to clean up, but it saves some hours since it can analyze a lot of law quickly and vaguely run a likelihood analysis.
Law is also in English. English is not formally logical. Frequently laws have subtle wording problems.
I assume people would very move quickly to get the speed limits changed to a number that represents the actual maximum speed you can go before receiving a ticket, rather than the vague percentage/luck combo that it is now.
For example, If there are so many laws that everyone is always doing something illegal all the time, then a corrupt cop can stop anyone he wants to at any time, because they've got to be guilty of something. It's a reason for types of deregulation, and keeping law slim in general.
When police have to overlook a large amount of activity that violates the law, something is wrong with the law.
I've heard this hypothesized as a means to remove legislative bloat, but where has it actually been implemented?
started as 2:1 https://www.theatlantic.com/business/archive/2017/01/trumps-...
6 months in, it was actually more like 16:1 https://www.washingtonexaminer.com/trump-kills-16-regulation...
I think I heard recently it's more like 24:1.
That someone should "make AI to do X" certainly isn't too much to ask.
Likewise, why can't just someone make a pill that gives us perfect memory and understanding of all laws on the books? Make the pill so you only have to swallow it once, and then your knowledge gets automatically updated. Make a pill that does all that, but is also very easy to make and has no side-effects, so there are no big costs or complications. Such a complex problem, and yet we both proposed very elegant solutions within minutes of thinking about it. Why isn't this this sorted already?
Law in its writing and interpretation is less deterministic, and quantity of statute is far less of a factor in its complexity than most in STEM realise.
Further, to parse law in a sense that it could be sensibly refactored, would be impossible with AI/ML as the field stands today, and I think will never be possible.
It's already accepted that everyone passing a border show photo ID and get their passage approved and recorded.
The facial recognition ship has long sailed for 1st world citizens thanks to how many pictures most of us have online, official photos for ID, etc, etc. The government can already map your face to your identity.
This particular program is probably more targeted at the non US residents crossing the border because the three letter agencies don't have access to as much good data and metadata on them. Because all the Mexican government agencies don't share ID photos with the three letter agencies the three letter agencies have to do it themselves hence facial recognition at borders.
As others have said I'd be very surprised if they weren't just trying to legitimize something they were already doing.
Photographing and storing data on millions of people without probably cause, sharing them with third parties (like Amazon) and keeping them forever, that's not just unnecessary, but also dangerous.
https://news.ycombinator.com/item?id=16327728