Schneier on Security
A blog covering security and security technology.
« New Internet Porn Scam |
| Marketing at the RSA Conference »
March 5, 2013
Technologies of Surveillance
It's a new day for the New York Police Department, with technology increasingly informing the way cops do their jobs. With innovation comes new possibilities but also new concerns.
For one, the NYPD is testing a new type of security apparatus that uses terahertz radiation to detect guns under clothing from a distance. As Police Commissioner Ray Kelly explained to the Daily News back in January, If something is obstructing the flow of that radiation -- a weapon, for example -- the device will highlight that object.
Ignore, for a moment, the glaring constitutional concerns, which make the stop-and-frisk debate pale in comparison: virtual strip-searching, evasion of probable cause, potential racial profiling. Organizations like the American Civil Liberties Union are all over those, even though their opposition probably won't make a difference. We're scared of both terrorism and crime, even as the risks decrease; and when we're scared, we're willing to give up all sorts of freedoms to assuage our fears. Often, the courts go along.
A more pressing question is the effectiveness of technologies that are supposed to make us safer. These include the NYPD's Domain Awareness System, developed by Microsoft, which aims to integrate massive quantities of data to alert cops when a crime may be taking place. Other innovations are surely in the pipeline, all promising to make the city safer. But are we being sold a bill of goods?
For example, press reports make the gun-detection machine look good. We see images from the camera that pretty clearly show a gun outlined under someone's clothing. From that, we can imagine how this technology can spot gun-toting criminals as they enter government buildings or terrorize neighborhoods. Given the right inputs, we naturally construct these stories in our heads. The technology seems like a good idea, we conclude.
The reality is that we reach these conclusions much in the same way we decide that, say, drinking Mountain Dew makes you look cool. These are, after all, the products of for-profit companies, pushed by vendors looking to make sales. As such, they're marketed no less aggressively than soda pop and deodorant. Those images of criminals with concealed weapons were carefully created both to demonstrate maximum effectiveness and push our fear buttons. These companies deliberately craft stories of their effectiveness, both through advertising and placement on television and movies, where police are often showed using high-powered tools to catch high-value targets with minimum complication.
The truth is that many of these technologies are nowhere near as reliable as claimed. They end up costing us gazillions of dollars and open the door for significant abuse. Of course, the vendors hope that by the time we realize this, they're too embedded in our security culture to be removed.
The current poster child for this sort of morass is the airport full-body scanner. Rushed into airports after the underwear bomber Umar Farouk Abdulmutallab nearly blew up a Northwest Airlines flight in 2009, they made us feel better, even though they don't work very well and, ironically, wouldn't have caught Abdulmutallab with his underwear bomb. Both the Transportation Security Administration and vendors repeatedly lied about their effectiveness, whether they stored images, and how safe they were. In January, finally, backscatter X-ray scanners were removed from airports because the company who made them couldn't sufficiently blur the images so they didn't show travelers naked. Now, only millimeter-wave full-body scanners remain.
Another example is closed-circuit television (CCTV) cameras. These have been marketed as a technological solution to both crime and understaffed police and security organizations. London, for example, is rife with them, and New York has plenty of its own. To many, it seems apparent that they make us safer, despite cries of Big Brother. The problem is that in study after study, researchers have concluded that they don't.
Counterterrorist data mining and fusion centers: nowhere near as useful as those selling the technologies claimed. It's the same with DNA testing and fingerprint technologies: both are far less accurate than most people believe. Even torture has been oversold as a security system -- this time by a government instead of a company -- despite decades of evidence that it doesn't work and makes us all less safe.
It's not that these technologies are totally useless. It's that they're expensive, and none of them is a panacea. Maybe there's a use for a terahertz radar, and maybe the benefits of the technology are worth the costs. But we should not forget that there's a profit motive at work, too.
An edited version of this essay, without links, appeared in the New York Daily News.
EDITED TO ADD (2/13): IBM's version massive data policing system is being tested in Rio de Jeneiro.
Posted on March 5, 2013 at 6:28 AM
• 33 Comments
To receive these entries once a month by e-mail, sign up for the Crypto-Gram Newsletter.
So when will the first person by shot for carrying a concealed spanner?
Way over compensation for threat. So, we had 911 and start looking for needles in haystacks -- and we are looking in the wrong places.
This all seems, ultimately, to me, to be about power. These people buying these products, implementing them, going that direction ultimately seem to be doing so because of lust of power.
There are likely money exchanges made.
But the *flaw*, I see as in their nature. They are not thinking matters through. Whatever power they may get from such moves, what does it matter? They will be dead in a flash. They live quickly and are gone.
They do not engage thinking faculties required for understanding other people. So, they end up hurting a lot of innocent people: even if that hurt is just inconvenience, invasion of privacy, spreading fear, and the like.
Ultimately, the thinking process and heart process involved there is identical to the kind involved in criminal actions, including terrorism.
They are simply not locked up for it, even if they are cursed quietly by everyone it negatively effects. And, this, in turn, causes people to consider the value of the authorities and laws even less. These authorities are making mockeries out of these things.
Give them a hundred years or a few more disasters and we will be right back to the culture of the dark ages. China. North Korea. Nazi Germany. Soviet Russia.
No lessons learned.
Terahertz radiation? Is that good for you???
Terahertz: hurts terror - obviously
The UK has a program called Crimewatch which frequently features CCTV footage which assists in the identification, prosecution and conviction of criminals. Whether it passes a cost benefit analysis vs some other form of policing is another matter. But it does prevent crime if only by ensuring more criminals are put behind bars and not free to reoffend.
Hey, you forgot one: the polygraph. (I wince every time I hear one called a lie detector, because it ain't and it don't.)
And of course it will trigger on those that can carry a gun concealed lawfully, including, I assume, off-duty cops, so one would think that indiscriminate use would have practical issues.
The problem with these scanners is it's an automated way to pull the trigger on an innocent person.
We have plenty of cases where an innocent person is accused of being armed and then shot. At the inquest there is all sorts of prevarication by the authorities and at the end of the day the officer that shot the person usually walks away and is bought a drink or ten by their colleagues.
So what happens now, a police officer gets told by a "gizzmo" that something is concealed on a person the officer pulls a gun his colleaggues pull their guns and like as not one of them is going to squease the trigger these days.
So what happens at the inquest the police officer his authority and union all blaim the technology and the fact that the deceased did something such as move in the wrong way or did not move in the right way when ordered to do so.
But no worries the technology will take the rap, only it won't because it will have cost to much to be decommisssioned so there will be a change in "operational proceadures" untill the next time it happens and the next time and the next...
Then of course there is the issue of innocent others who just happen to be in the wrong place at the wrong time and catch the sstrays or through and throughs.
Whilst I won't say there is no place in the police for technology, I will say it should not replace the human evaluation process in any way.
Gun-spotting radar -- what a great tool for safety-sensitive robbers to have! Maybe a job for the getaway driver while they wait. And another great way to spot undercover police officers and those pesky gun-toting federal air marshals.
On a more serious note, does anyone have any doubt that law enforcement will gleefully scan every individual and group of "gang" youth of color (and other people they don't like -- gays, Muslims, politicals)? And what about the many-times mugged and/or raped person who carries illegally? Maybe the next Abner Louima loses his life because his privacy-protecting metalized wallet is mistaken for a gun.
In line with the whole theme of "if we go looking for needles in haystacks, we might miss the Magnum Research BFR that could hurt a lot worse than the needle" At an incident response training school sponsored by Mandiant about 5 years ago, they mentioned how few incidents were detected from classical Network IDS/IPS (percentage in the single digits) and classical A/V (very low double digits). More incidents were discovered by end users or could be traced to notification by external parties. Admittedly, network IPS has come a long way in the last 5 year, and people are looking at egress streams, which was rare 5 years ago. But it goes back to the "If all you have is a needle finder, because someone sat on one, and it made both CNN and FOX, then every problem looks like a needle." Or maybe the expression had to do with hammers and nails. ;-)
Not just the cost of technology/efficacy...
My back of napkin calculation for security related to terrorism is based on $10 trillion dollars to FYI 2015 and an estimated 8500 terrorist threats eliminated/repressed yields a raw dollor cost of over...
$11 Billion dollars per terrorist...
Now that's value!
And how many died in the process executing this bargain...
@name.withheld.for.obvious.reasons: Using the standard of 10 million dollars per person that means each of those stopped threats would need to be on the order of 9/11 for it to be worth it...
As a technology seller I'm all for this, whats better than selling fairly useless technology. Once you perfect the technology you can make them pay again for upgrades, just look at the TSA to understand how the game is played.
Interestingly once you perfect the technology it is VERY likely that it will be ruled illegal for the reasons Bruce outlined, so fantastic you now get paid to decommission (safely) and disposal of in a manner that the technology does not fall into enemy hands.
As seller, what's not to like about this deal?
Now my biggest concern is for the possible side effects of THz radiation. Truth is nobody knows what possible health effects are of using these frequencies at relatively high power levels.
But as for the technology I love it, better still I love the easy counter technology because jamming this is trivial, hence the need for upgrading.
As seller, what's not to like about this deal
Well a lot actually depending on your view point.
First off as a technology step in it's own right it's likely to be a technology cul-de-sac and if it is short lived with high startup costs. One side effect of this is the labour costs and other fixed costs unless you find a way to offset/outsource them.
Secondly as we are starting to find out contracts are not fair, open, or honest and appear to only get given to "those with the right contacts". As we know maintaining those contacts can be very expensive not just short term but long term.
Thirdly as you noted at some point the merry go round has to stop and the piper payed... The question is what has changed in the mean time, a look at the fortunes of Ken Lay and others at Enron and other similar organisations indicates that Joe Public who feels "scr3w3d by sub prime" is looking for payback, the question is who are the press going to "whip up a 5h1t storm" for.
Now for the well connected usually they can buy off the press one way or another, including dishing the dirt on a larger scalp.
It's easy to see addressing this is "non profitable" activity and to be quite honest sailing rather closer to the wind than many would care to go but as some would say "no pain no gain".
You will also probably need to consider some interesting financial engineering due to "product liability". As you note,
Now my biggest concern is for the possible side effects of THz radiation. Truth is nobody knows what possible health effects are...
There was a time when asbestos was the "wonder kid of the age" and it went into everything including as a filler in plastics used for eating utensils and cookware.
Then people started to get ill and die in unpleasent and unnatural ways, and whilst the fleet of foot managed to get out in time many did not and the price just kept rising and rising. And some lawyers became effectivly "Kings of Torts". And guess what the lawyers are getting smarter than those running these companies and they know how to dig in ways that were unimaginable just a few years ago.
As with coruption, the public tolerance to product liability is changing, it happens faster and harder.
So whilst there are some healthy initial profits to be made the whole model is quite risky these days.
Much of what Bruce says here is also applicable to health care. A great number of technological (mostly IT) systems, also costing gazillions, are marketed as silver bullet-solutions for growing problems in the health sector. Desperate politicians like to believe these claims. But EHR / EMR systems, for example, promised to give better care and lower costs, actually don’t deliver neither. Potential for abuse – for example, how easy it is for unauthorized people to access your electronic medical records – is typically downplayed. Or, even more IT technology is thrown at the problems, creating even more inefficiency. And then the systems quickly become to embedded (in the health care system, in this case) to remove.
When the airport scanners came in, there was an article about them in my local paper and I sent an e-mail to the reporter, saying that they should have asked "Have these been independently tested, or will they be?" She wrote back, saying that she asked that question and was told in no uncertain terms that no independent testing of the machines would be permitted.
I took that as an admission that the machines don't work, and that the TSA didn't care if they worked, they just want to funnel tax money to their friends who make machines.
Let's see if the people making these scanners are any more willing to consider independent testing. I'm betting that they aren't, and that the government won't be either.
"You will also probably need to consider some interesting financial engineering due to "product liability". As you note,...."
these are valid big company concerns but as a startup you either find customers or you go out of business very quickly. Here is a product with a moderate level of technical difficulty, which will burst onto the scene creating reasonable volume and great unit prices (for a while). Now if it proves to be detrimental to health then thats tomorrows battle, I can assure you the ambulance chasers only ever chase the successful, so that's a risk of succeeding.
To be 100% honest it is not a risk a successful startup ever faces because IF the product is that successful then the startup company will be bought out by a larger politically connected "defense systems" type company, which assumes the liability.
Today's 20nm CMOS has Ft's above 300Ghz so it is possible to work directly with THz frequencies, power dissipation is a problem for sure but it is technically possible to design these RX/Tx products directly into low cost CMOS processes.
There are a number of interface /antenna problems to be solved but when the basic technology is there a way will be found to use this capability.
As for the technology being a dead end, i hope so but realistically I don't think so. There is too much value in the eyes of the self righteous to give up this capability. From my own experience most military, civil defense and LEO personnel are extremely self righteous, they can see nothing but good things resulting from their access to better more complete picture of the situation. The fact that real information gets buried in the noise of technology is completely lost on them. Meaning I have no doubts about the success of this technology.
Unfortunately the information asymmetry this sort of tool creates will also cause criminals to act first and wonder later if the cop has or does not have this capability.
The issue with start ups outside of what is in effect "tax fleacing" is often Angels who still have the "burn rate" mentality.
The simple idea being to be either first to market or a little behind the bleeding edge to get market capture. Simply by throwing money into the company.
The result is usually an engineeering of an investor bubble which can be good for the Angels but less so for all but a few technology share holders.
I've been out of "tax fleacing" development for a number of years (thankfully) and more recently out of FMCE (it's a "Red Queens Race" these days).
But I still keep a watching eye and I've noticed that "terror" keeps inflating the bubble in both sectors for well past what would normally cause a market collapse. And for obvious reasons this is a concern. What I cann't get a real grip on and I suspect many others cannot either is why the bubble has not popped yet. I positivly do not believe that any market can survive such an investor bubble unless it produces real value at some point to actually offset the investment inflation.
Various supposadly blue chip High-Tec companies are begining to run into investor bubbles take a look at Apple's share price / dividend returns of late and their apparent determination to enter a suicide pact with various other companies. Such behaviour is pushing up their burn rate but I don't see any commensurate offset from the market...
As has often been said "time will tell" and "the study of history gives cautionary warnings" but I realy don't see any evidence of the latter currently...
> they're too embedded in our security culture to be removed.
I was at the Colorado state capitol yesterday, hoping to speak out against the gun control bills before the legislature. (I didn't get a chance, the lines were so long).
One thing I wanted to ask our elected representatives was if these measures -- which they claim will make us safer -- become law, will they remove the metal detectors from the public entrance to the building we were in?
The surveillance technology I am watching for is little civilian quadcopters with cameras. DIYers will come up with really hard to jam or spoof radio controls, and stream live video to safely remote recorders . . .
The legal problem seems to be that laws don't take into account things that simply weren't possible when the law was written (how could they, after all?). The writers of the fourth amendment couldn't possibly anticipate the issue of technology that can see through clothing.
And maybe the answer is to try to translate this new technology into the traditional method of accomplishing the same thing, then ask if THAT would be legal. NYPD has a device that can see a gun under my clothing and they want to know if they can use it? Would an NYPD officer be allowed to go around randomly frisking people? In fact the courts have specifically ruled that they can't, a new technology to do the same thing should make no difference.
@Neil in Chicago: That raises some interesting possibilities. Police officers typically aren't THAT numerous compared to the population of any given area, and squad cars could reasonably be tracked from the air. How difficult and expensive would it be for a group of hobbiests to run some drones able to track, in real time, EVERY police car in their city? Or how about any arbitrary group of people...say, the local elected officials. Maybe non-trivial, but likely not incredibly complicated either.
I'm not for a second suggesting it would be a good thing for anyone to do that, in fact I think it would be bad for any number of reasons. But the legal decisions surrounding new technologies need to look past the good uses and consider how they might be used negatively. Our laws shouldn't be based on assuming everyone with a given technology is a good guy.
" How difficult and expensive would it be for a group of hobbiests to run some drones able to track, in real time, EVERY police car in their city? "
Or just cheap wireless transmitters? Or a modification to the car's own electronics? Or by tracking the car via its radio or data uplinks? These seem a tad less noticeable and labor intensive than drones. GPS, cellular and otherwise wireless tracking are also proven techniques for tracking vehicles inexpensively, whereas drones + computer vision is new territory for amateurs.
Looks like there might be a brisk market for "drone catchers"...it's a no-brainer to capture drones and re-program/re-target their systems applications. Saves a whole lot of money.
One method is a kite-based drone-dragger that uses a drop net released by a trigger on the kite. Balloons are also an obvious tool to engage all kinds of drones--the advantage is power/weight/capability ratio using a balloon-based system. Ballon-based systems are also useful if EW is the game at play...
I don't know why, but the backscatter x-ray systems are still in use almost all the time in San Diego, where I must frequently deal with the annoyance of opting out. I haven't seen any visible progress toward the stated goal of removing them.
Technology is neutral. Application/purpose of usage assings + or - to person in charge of usage.
Can anybody elaborate on milivision scanners (active and passive) working in the same manner?
What is the cost of ThZ scanner?
By the way, it'll detect not only gun, but badge, hidden wire for recording of conversation, in particular hidden on nice-looking honey-trap like lady, i.e. Nikita is done with that.
About usage by police: all discovered facts (using any technology in particular)could be used in two ways:
as evidence for the future prosecution or just for routine police work.
In latter case the purpose of collecting primary facts is to move to collection of former. Those are basics.
I notice that we're seeing a lot of "marketing" language in play here. In case folks missed it, "Terahertz Radar" and "Millimeter Wave Detection" are the same things. Note that:
1Thz has a wavelength of 0.3mm
So this is just a variation of same "naked body scanner at the airport" that everyone was in a big uproar about. I also wonder how well this works on the street, as the ones at the airport seem to depend on people being still/etc.
I keep thinking, this would be an interesting thing to take along to an NRA convention. Complete with snipers on rooftops and all the paraphenalia of a counterterrorism lunatic.
Thank you for your posting. That technology was developed even before 9/11 for through the wall vision as well + special device working in concert to assess living status of the person behind that wall, meaning privacy issue is not when you just not naked in the oublic place only, but when you behind the wall of your own home. Question: is metal foil blocking through the cloth/wall vision or Faradey cage only?
Thanks again for your posting.
On the CCTV aspect see the article in Slate, with the headline 'Chicago Installed Thousands of Cameras on its Rail Platforms. Crime Jumped by 21 Percent'. Link: http://www.slate.com/blogs/crime/2013/02/26/...
One wonders if every law enforcement official was motivated and professional what difference that would make.
Here in the UK the police have steadily withdrawn from the streets, mainly to complete paperwork and fulfil desk-based investigations.
Sounds like CPD have more problems than high demand.
Great post. Brilliant mind....and here we are hoping you will join us at http://veillance.me - an IEEE SSIT event specifically examining these issues at hand.
Looking at the description, I can't tell if it's reading natural IR emissions, or irradiating the target and reading the reflection. If it's the former, wouldn't just keeping the gun at ~98.6F defeat the scanner?
Schneier.com is a personal website. Opinions expressed are not necessarily those of BT.