I think the big thing people forget when they automate systems is that they make them too strict. As in you're always expected to fit within the parameters and pattern of behaviour that the system was built with. It tolerates no deviation.
However as with all human things, it is better and less trouble to be able to take shortcuts rather than have to stick to the rules 100%. For example people crossing the road in the middle of the night with not a car in sight. People reversing down the road a little bit to turn or get a parking spot. Being able to cross the centre green patch on a high way at night because there won't be a u-turn available for many kms. The issues with identical twins and not being able to get IDs (basically the inability to override a system).
These sort of issues help remove stress and simplify things in our daily lives, and help make the world function.
What has that got to do with the article? Facial recognition isn't crime detection...it's recognising someone who has an arrest warrant (or in this case, perhaps also checking against a list of banned hooligans), i.e. it either is the person or it isn't.
My comment was on people putting too much trust into computerised systems, of which facial recognition is one. It also referenced comments other people had made here for when the rules and programming in such a computerised system broke down in the real world.
In the case of facial recognition there was examples of twins being rejected for IDs because the system thought it was the same person applying under different names.
I'd also be dubious as how well it can identify people based on grainy and low resolution camera footage.
Let me wave my hand around a crystal ball and peer into our future...
"Cameras are everywhere and record every crime. Somewhere in the city a man commits a murder. Computers register the crime. Later that same man steps into a self-driving car. The man explains to the car that he wants to travel to the other side of the city. The car doors lock and the vehicle takes off. But the car is not going to where he thinks. Slowly the man realizes that he is being taken directly to the nearest police precinct for booking. Desperately the man tries to exit the car but the doors are locked. He lays down in the seat and frantically kicks at the windshield but there is no time. The car has been given emergency priority and begins traveling over 90 mph. The precious seconds tick by...
Back at the precinct the cop takes another bit of his doughnut and reminisces about the time when policing was not just a desk job."
Some AI is interpreting all camera around the country in real-time. A single government employee is adjusting parameters according to what the current elected officials deem important. There is, however, extra CPU time dedicated to cameras monitoring political opponents.
You'd probably like the show Person of Interest written by John Nolan -- Christopher Nolan's (writer and director of Dark Knight trilogy, Intersteller, Inception, etc.) brother.
It loosely follows this premise of multiple AGIs going head-to-head against each other as one is used for "good" and the other for "evil" by their respective creators.
Hehe :) Seconding recommendation for Person of Interest. Jonathan Nolan then went on to create the Westworld HBO series. These two brothers are really talented, it runs in the family.
Then I think you found a great startup idea. Create the AI that warns you when you are about to inadvertently commit a crime, thereby helping you navigate the complex and incomprehensible legal system. Investors will line up to give you money because they will know that everyone will need your patented AI to avoid jail
Or a new type of actor for hire that gets people to inadvertently commit crimes on cameras. Same reasons as 10,000 years ago: angry spouses, jealousy, political assassination etc.
1) Ubiquitous self-driving vehicles + facial recognition + cellphone gps tracking = plausibly deniable targetted killings on a massive scale.
In 2030 when the AI watching you on all the security cams decide your behaviour pattern skew two standard deviations toward terroristlike, your road accident rates go up.
The AI-driven car insurance companies are the first to catch on. When your car premiums shoot skyhigh for no reason, time to stop doing whatever you were doing.
2) 2030. Some researchers in a big prestigious university decides to do a study on SDC accident rates. They find that cars from a particular major maker have a tendency to hit women and black people more often. Results are published in a major journal, media has a field day.
The company apologizes profusely and promises to fix its racist mysogynist ways. Their software-engs are left scraching their heads because they can't find the bug. Everything seems to be working... except accidents still skew towards politically sensitive demographics.
In desperation the company decides to very slightly decrease the threshold for hitting non-black non-women, evening the numbers. This patch is combined with a major service release which improves accident rates across the board, so that the overall accident rate stays the same.
Six months later the journal publishes a very small correction for the original paper that fixes a few bugs in its statistics. The correction is behind a paywall. Nobody reads it.
sure, then it all falls on its face due to a stupid bug. I'm currently writing a book on cases where bugs caused serious damage or harm to people and it's amazing how many stupid assumptions people make believing that tech is always right. e.g. Georgia Department of Driver Services couldn't handle photo recognition for twins -- their fraud detection systems insisted it was the same person and rejected driving license test applications as potential fraud (same person applying twice under different names). BBC journalists managed to trick HSBC voice authentication (again using twins). Massachusetts Registry of Motor Vehicles falsely identified a person by matching his drivers license photo against an anti-terrorism database and revoked his license, wouldn't even tell him over the phone why. It's just a matter of time before some really serious mess kicks off because a computer wrongly recognised something.
But all of this was happening before, and it's still happening all the time: people getting arrested (some even convicted) just because they looked like someone else, bureaucrats refusing to issue documents to people because of some paperwork mistake, etc. My uncle had at one point to prove to police that he's still alive in order to get a new driving license, because someone with the similar ID number died, and they accidentally marked him as deceased. These kinds of errors always existed, and probably will always exist, just that before they were always caused by human factor, and now we have technology related errors too.
I've been the wrong side of a case of mistaken identity by a government department. It took a very long time to get it sorted out, despite the obvious impossibility of the situation their system claimed we were in, and while we aren't talking life and death, the consequences could have been quite serious.
The most frustrating thing was trying to get anyone to acknowledge the possibility of the error in the first place. Even some years ago when this happened, there was so much blind trust in the almighty database that it was basically dumb luck that I eventually managed to find someone willing to take a deeper look without just hanging up the phone or otherwise cutting me off.
The biggest underlying problem, apart from the original error that caused the whole situation, seemed to be that despite my complaints there was no sort of formal case opened to deal with the problem, nor any clear set of records kept, nor any specific individual assigned to resolve the matter. The whole system simply didn't acknowledge the possibility of a mistake or provide any process for correcting it.
At least in my case the problem was obvious to me. With modern systems making automated decisions about anything from mortgage applications to who goes on a secret no-fly list, there seems to be a very real danger of innocent people suffering very serious problems as a result of these errors but having no way to even identify the cause of the problem to start trying to get it fixed.
eg. in australia, they launched the derided 'robo-debt' system that sent people notices about potential benefit fraud charges, and the harassment of people caused senate investigations in several states. turns out the new system is as correct/wrong as the manual process (20% error rate), but it's now able to process several order magnitude more cases, so effectively floods downstream workflows (claims, complaints, courts) that were not ready for such an increase in demand.
Or he commits the crime of unauthorized copying of software. Or building unlocked computers. Or helping people encrypt their data with secure, government-proof encryption.
That paragraph seems to be a paraphrase (intentionally or not) of something I read a while ago, I'll be damned if I can remember what though.
I vaguely remember it being a girl being arrested by an automated car and her having to hack her way out, maybe while going over a bridge. Damn my shoddy memory!
I think it might have been something by William Gibson? I believe she had to cut through the seatbelt and then call a friend to step out in front of the car causing it to career off the bridge and into the water, at which point the locks would release. (The ethics calculation deemed it preferable that she escape rather than perish.)
given all the different moving parts described in this scenario, the bureaucratic challenges alone make me immediately file this in the jetpacks and flying cars category of sci-fi/tech. the changes that would have to happen in government to make this work would be the real paradigm shift. what a world that would be.
The problem is trying to "automate" the justice system, and in the process, accidentally creating a dystopian dictator's fantasy.
Mass surveillance provides unparalleled power to those who control it. It's unprecedented in all of human history. I don't believe humanity knows how to safely handle such power at this point.
One of the most fundamental concerns is the fact that, once collected, the data is vulnerable to being misappropriated or weaponized for the entirety of its existence, possibly for many decades or longer. By that point, laws may have changed significantly and the entire political climate could be drastically different.
Imagine being automatically put on a special list and thus denied certain basic priveleges due to "questionable" posts you made in an online forum 15 years ago. We can imagine a wide variety of vague and controversial definitions of "questionable content" for this thought experiment.
But the police is explicitly _not_ the justice system.
Courts are of course complicated, and I imagine that police would abuse such a network to arrest people without means to fight back... but at the end of the day I can at least hope to be in front of a judge. And it's not like the courts are the ones doing the dystopian investigating.
First one who said "the innocent have nothing to fear" was some Maximilien François Marie Isidore de Robespierre. He was eventually executed. Turns out, it depends a lot on who decides your innocence, and not a lot on you doing nothing wrong.
In other words, systems set up against people doing bad things are also perfect for framing people inconvenient to the operators.
It's wonderful, when the punishment for murder is a Norwegian prison, terrible when the punishment for shoplifting is ISIS chops off your hand.
There's a broad region in between those two poles, where some cases are merely somewhat awful and not completely horrible, and other cases where I guess it's kind of okay, but not that great. The rest of the time, it's an apathetic, civil neither.
The majority of criminals are good people who deserve no punishment. The majority of criminals who deserve punishment don't receive it. The law has grown to the point where perfectly moral and reasonable behavior is criminal. Automating a authoritarian system of enslavement is not a good thing when it has become so divorced from morals or ethics.
If anything, we need to slow down the justice system. Remove plea deals. For every charge to go before a jury. Force the DA to focus only on the bad crimes and let anyone with minor crimes off with no charges.
A sad fact of life, albeit based on life experience and anecdote, being young and female will mean you are being watched and followed even though you have never committed a crime.
Imagine a world with no crime. You'd probably not want to live in one, unless you're somehow very eager to do nothing but exactly what the government says you may, with the entire infrastructure preventing you from even attempting to do otherwise.
Hopefully at this point in the future humans will be augmented and have tools to circumvent this sort of thing. Confuse the AIs, jam the cameras, whatever.
So when a crime happens in public: in addition to a dump of all cellphones in the area they'll also have access to all scanned faces and license plates (via lidar) to reconstruct what happened. Plus potentially an always-on eye in the sky via drones/planes which allows them to "rewind" peoples and cars movements up until that point. Pretty scary to think how easy their jobs will be.
This already exists. I posted links in response to the OP referencing a RadioLab episode where precisely this occurred over Fallujah and helped to track down suicide bombers/bomb makers.
The'll have instant connection to any CCTV facing public places (including hotel corridors) and any dash cam. After all, all the videos will be uploaded to data centers via 7G.
Identification of humans, dogs, car plates will be in real time plus the identities will be logged for 5 years.
The CCTVs will record in PetaHD, enough for fingerprint identification from distance of 120 meters.
All cars will have black boxes with video and sound recording. Illegal turns will be autofined (if you were able to turn illegally, because most of the time your car won't allow you).
Don't forget logged local wifi, bluetooth and NFC connection records, and the sum total of any geotagged images or videos uploaded around that time and place.
Suspicion will be auto-assigned to people who are detected to have phones (via video analyzed pocket bulge or historic behavior) but which are not emitting signals due for example to 'flight mode', or those who lend NFC identities, which may become illegal or against terms of service.
> Plus potentially an always-on eye in the sky via drones/planes
I'm convinced that this will be a "bonus feature" if/when Google/Facebook/SpaceX get their balloons/drones/satellites off the ground. I don't want to live in that future.
That would make a cool sci-fi scenario. Man robs bank... wait, who would still use cash in the future? A man kills another, but the computers show him at home, waking up late, travelling to work, logging on his work PC. Except, the killed man was an owner of one of the encryption keys for the global metadata logging/surveillance system, and the criminal managed to kill that man and steal the key, altering all the computer logs to cover his tracks.
> In the same statement, the police force turns a little more towards pre-crime. South Wales police and crime commissioner Alun Michael said: "Our approach to policing is very much centred upon early intervention and prompt, positive action; the introduction of facial recognition helps to support these aims by allowing us to identify vulnerability, challenge perpetrators, and reduce instances of offending within environments where the technology is deployed."
As spooky as automatic facial recognition is for finding and arresting criminals, it's like they're parodying themselves with this statement. You couldn't sound more evil if you tried.
I think the real battle is in Judicial reforms. Honestly we are at a stage where the face recpgnition technology is just too easy to deploy. [1, 2]
There are also several legitimate use cases, e.g. Amber alerts, Public right to record, etc. The society can delay dealing with technology for only so long. Over the next decade we will see huge adoption of Autonomous cars, robots, cheap drones essentially "Mainstreaming" of surveillance technology. At this point the legal system would have to decide a sane framework to resolve the challenges posed by technology. However given the paralysis in Washington, and inability to deal with even the most basic questions of algorithms and law (No-fly lists being the prime example) I am not too hopeful.
Done by a government with proper legal oversight it certainly has its place and easily will be a net positive for society. Though lawmakers seem to be oblivious to tech advances moving far quicker than they can, don't hold much hope they can keep up.
Used by corporations it will be frustrating and take privacy intrusions to the next level. In the end though, democracies will survive fine.
The real issue is tyrannical regimes getting their hands on facial recognition, it's a terrifying prospect and likely to be used with great effect.
Last night I watched the 'Total Recall' remake with he kids. I couldn't help making connections between the film and the way technological policing is progressing. I keep noticing that in dystopian sci-fi films there are far too many parallels with the UK. I should stop watching sci-fi and finish some coding...
Meanwhile, back in the real world, I spent an hour waiting to speak to someone in the police the other day only for the guy on the desk to say he needed another 30-45 minutes with the lady at the desk and there were no other staff. So, if you really do want to report a crime or some criminal behaviour that has got to the stage where a little bit of community work by the police would help, good luck to you. Maybe take an afternoon off work so you will have time for the bureaucracy.
We are a long way from 'you have twenty seconds to comply' policing.
It seems to me that the danger is having no time for anything but 'you have twenty seconds to comply' policing.
Just look at the comments made by serving UK police officers in light of the recent high-profile attacks, particularly about the relative lack of intelligence now that there are so few officers assigned to basic community policing and everything has become about reaction rather than relationships.
Compare and contrast those views with the recent comments by Theresa May about how the problem wasn't really caused by a sharp reduction in the number of police officers but instead must be fixed by regulating everything and throwing human rights laws out the window if they happen to be inconvenient.
My experience was definitely of this, the problem that you cite.
But as you mention, police, despite all accusations to the contrary, are not stupid. They know what policing is and it is about communities and people rather than enforcing the edicts of a given political party. It is strange how police are always politically controlled, maybe in some do-gooding place like Norway they are allowed to be professional without interference.
Along with 'twenty seconds to comply...' there is this instant execution by police that happens. Nobody is shedding a tear for the three terrorists shot 8 minutes into their killing spree, however, would the police have been able to apprehend them if they only had their batons?!? This was the situation in the UK until recently, if you 'went postal' in some regional town it would take a while before officers with guns would arrive. You would have hours to escape bullets. But now? 8 minutes.
I also imagine a world where no terrorist gets tried in court. Death awaits them before they get that far and the quaint notion of justice before a jury goes by the by.
You’re not going to get an 8 minute response anywhere but in the major city centres, which have armed units on 24/7 standby.
I think those units have been around in London for decades at this point. Probably ever since the ring of steel was installed as a response to the IRA bombing campaigns of the 70s/80s.
Exactly. Borough market is right next to London Bridge station, Guy's Hospital and the Shard, all of which are obvious places to park rapid response units.
Similarly the area of Westminster around Parliament and the Palace will be completely covered 24/7, extending out towards the nearby embassies and Whitehall civil service buildings.
I was recently the victim of a random assault. Police refused to come to the scene ("no patrols were available"). The nearest police station was closed ("it was too late"). The second nearest police station was all but abandoned, so I had to wait around for 45 minutes. When a policeman did show up to take my statement, he was uninterested and spent most of his time avoiding my gaze. After the statement was done, he told me not to bother pressing charges because no one had the time to look at the camera footage in the area.
However as with all human things, it is better and less trouble to be able to take shortcuts rather than have to stick to the rules 100%. For example people crossing the road in the middle of the night with not a car in sight. People reversing down the road a little bit to turn or get a parking spot. Being able to cross the centre green patch on a high way at night because there won't be a u-turn available for many kms. The issues with identical twins and not being able to get IDs (basically the inability to override a system).
These sort of issues help remove stress and simplify things in our daily lives, and help make the world function.