2084 — Dodging a dystopian future

2084 is a 100 years after the events in George Orwell’s 1984 and far enough in our own future that in all likelihood I will get some stuff wrong trying to predict the future. Also, in all likelihood I will be either dead or be benefiting from future progress in medical research as I should be 110 by then (I have good hopes).

George Orwell presented a very bleak and dystopian vision of what the future would look like that was published in 1949. At the beginning of the cold war and right after WW II this was a message that resonated well. But ultimately things did not quite play out as he thought it would and he both underestimated what would be possible technically and overestimated how people abuse these capabilities.

In this article I want to take a good honest look at where we are with mass surveillance technology and extrapolate on some clear technical trends that are driving that forward. What we have today is in many ways scarier than anything Orwell imagined; yet we maintain reasonable levels of freedom and are arguably doing quite well on average. Is that going to stay true going forward?

In the past two decades we’ve deployed billions of internet connected mobile phones, computers and other devices, and we’ve become dependent on the internet based services that run on them. As part of this process, the amount of information available to governments and other parties to track, spy on, and control us has also increased.

A lot of people are rightfully worried about all this. A lot more of them are blissfully ignorant of just how easy it is to track their every move. It’s gotten to the point where the only safe assumption is that you are under some form of surveillance at any point. While this is not necessarily true for most, to assume otherwise has a non zero chance of being misguided.

So, lets look at our hypothetical 2084 world, 67 years from now. Technology has made all the leaps we envisioned and probably exceeded our wildest expectations. What does that mean in terms of technical capability for the surveillance of people, legislation and practicalities around this, and possible defense/counter measures individuals could take?

This is roughly the question that I started pondering one Sunday morning some time ago. It’s easy to predict the negative Orwellian outcome and frankly a bit of a cliche. George Orwell got some things wrong, and do things really have to end so badly? Besides, we’re living the dream (or nightmare?) in terms of what last century dystopians imagined and arguably things aren’t that bad. Relatively speaking more people than ever enjoy quite pleasant lifestyles and are unaffected by things like war, famine, disease and other forms of unpleasantness. Also there are more of us than ever and quite a few of us enjoy freedom of speech. At least I don’t feel particularly oppressed writing this. So, overall, things are arguably better on average for most. Of course not all of us are that lucky but generally speaking the dystopian future has failed to materialize for most of us. But if you live in e.g. North Korea, you might disagree and be somewhat reluctant to express that opinion.

Ubiquitous surveillance

I believe that the logical result of expected (and probably some unexpected) technical improvements will result in what for lack of a better term is ubiquitous — impossible to escape — planet wide surveillance of everything and everyone. Or ubiquitous surveillance for short. This part is easy to predict and is already becoming a reality. Just imagine more of what we have today and it’s easy to see where stuff is going: surveillance is going to be everywhere, all the time. There will be insanely fast computers putting all the dots together and lots of technology providing plenty of little dots to put together. By ubiquitous I mean that you literally can’t escape it: everywhere you go there will be some form of surveillance technology in place. The resulting data will be stored more or less indefinitely as well as analyzed both in real time and after the fact.

The reason this is easy to predict is because we everything we need for this already exists. All we need is more and better versions of what we have, at a lower cost, and maybe with some improvements here and there. As an engineer predicting that this is our future is a no brainer since most of these things are evolving on a non linear scale and have done so for some time. All I’m predicting is that that won’t stop anytime soon. Technically I’m just stating the obvious.

You can’t hide

Just ask your self how far you are away from a mobile phone that is turned on at any point in your day. It doesn’t have to be your own phone. Any phone will do. Now imagine that somebody is able to turn on the microphone and camera remotely and listen in. This part is not science fiction. Now imagine that the microphone is on all the time and that some AI is analyzing everything it hears. This too is not science fiction. Google is listening for the magic words “OK Google” 24x7 on hundreds of millions of phones. Now imagine all this data aggregated and analyzed in real time, all the time. The NSA already has some capability to do this but but they have probably not quite yet reached the point where this becomes a “lets turn it on for an entire country” kind of thing. But it is not a stretch of the imagination that they’ll eventually be capable of doing that technically.

The technical feasibility of ubiquitous surveillance is such that you have to assume it will happen sooner rather than later. Think trillions of cameras, microphones, all sorts of sensors, etc. permanently turned on and streaming data all the time. Then consider the vast networking and computing infrastructure deployed by the likes of Google, the NSA, Russia, China, etc.

The amount of video stored by Youtube is measured in dozens to hundreds of hours of video per minute: every minute, hundreds of hours of video is coming in. Imagine that you are recording somebody 24x7 in 4K quality. How many hours per minute would you need to keep track of 10 billion people. Assuming the current state of the art of 100 hours per minute for youtube, or 6000 people minutes, that means youtube needs to ‘slightly’ increase their capacity to about 160 million times as much (for 10 billion people) to be able to track 10 billion people. That sounds like a lot but bearing in mind storage capacity has been growing exponentially, this is doable.

Storage: a numbers game

This little blurb comes from a data storage company, so you should probably take it with a grain of salt:

From 2005 to 2020, the digital universe will grow by a factor of 300, from 130 exabytes to 40,000 exabytes, or 40 trillion gigabytes (more than 5,200 gigabytes for every man, woman, and child in 2020). From now until 2020, the digital universe will about double every two years.

But just extrapolating from those trends we can deduce a few interesting things about 2084. According to the above we are a bit under 2 doublings away from 40K). So lets assume 10K exabytes today. I suspect this may be overly pessimistic but it should be ballpark accurate. A factor 2 is not that much when things are exponential. 2084 is 67 years away, or about 33 doublings of our storage capacity if the trend above continues. Lets assume it does.

Since I studied computer science, I’ll be extra pessimistic and take away one doubling so we can pretend it’s 2 to the power of 32. Or 4.2 billion, which is a nice number to work with. You might recognize that number as the available address space of old 32 bit PCs. So, whenever you read 2084, just think 4.2 billion times whatever we have today.

1 exabyte is 2 to the power of 60 bytes and if you know your high-school math, you’ll know that we can simply add the exponentials when multiplying. So we’re talking 2 to the power of 60+32=92 times 40000 exabytes. 40000 would be roughly ballpark 2¹⁵ (32768). So lets say the worlds storage capacity comes in at 2 to the power of 60+32+15=107 bytes. That’s about 1.6E32 bytes in the scientific notation, aka. a really long number with about 32 digits in decimal. That’s a lot of bytes in other words.

To put this in perspective. 1 minute of of 4K video on a smart phone is about 400MB. So, we’ll cross the mark where we can store the 80*365*24*60=42 million minutes of 4K video that make up the average human life long before 2084. A human life’s worth of 4K video is about 1.6E16 bytes. 1 exabyte is a mere 1E18. So you need around 0.01 (1E-2) exabyte to store a human life’s worth of video in 4K quality.

That is ballpark what you should be carrying around in your pocket in just a few decades. Lets do the math for that. Today, 32 GB is common to find on a smart phone. Or in exabytes that is 3.2E-8 bytes. We need 1.6E-2, so that’s only a few doublings away: 500 million to be exact. Or 2 to the power of 19 roughly, which means 38 years from now. By 2055 you’ll carry enough capacity to store your entire life in 4K quality on whatever you carry in your pocket by then. Whether you will is of course a different question but it is kind of mind boggling to imagine you could.

Analytics and AI

Companies like Google are not merely storing data, they also do things as face recognition, geo tagging, and using AI to recognize things in video streams. After all they are a search company and they don’t just store data but also meta data. Computer scientists are hard at work ramping up the capability to abstract from raw data and reports of impressive breakthroughs on that front are regular as clockwork. Computers can recognize people, objects, buildings, etc from raw photo and video footage today with high confidence. So that means by 2084 not only will we have the ability to store the video captured lives of each individual on this planet many times over in whatever ultra high resolution is common, that video will also be cross indexed against rich metadata that will be searchable in real time.

In practical terms this means it will be possible to track the movement of all individuals on this planet 24x7 for the duration of their life. Any second of footage anywhere will be analyzed, cross referenced, aggregated, and catalogued. The resulting metadata will be stored indefinitely. There will be a detailed record of anywhere you went, who you met, etc.

Identity: no more passports needed

This is where things get interesting (or scary depending on your point of view) because so far people have always been able to hide and disappear for a while. But how do you hide in a world where anything with a camera can instantly recognize you and tell others about the fact that it saw you. How would that work when just about every public space already has security cameras? When most people are already carrying smart phones and people are contemplating putting cameras in their glasses so they can experience augmented reality? That’s today, not 2084.

Not being able to hide in time and space is highly disruptive. It will be impossible to dodge or escape countless of machines equipped to track and recognize you. If you exist anywhere, it will be on record, permanently. So one interesting consequence: you won’t need a passport to prove who you are; machines will just know from looking at you. Nor will you be able to use a forged passport to pretend that you are someone else because bio metrics and other features tie you to piles of archived data that prove who you are.

The same applies to credit cards. All a credit card is is a rather convoluted way of proving you are authorized to approve financial transactions associated with some account. One of the lesser James Bond movies (GoldenEye) featured Pierce Brosnan walking into a very expensive hotel wearing nothing whatsoever after a unlikely escape from captivity & months of torture and getting his “usual suite” based on the notion that the manager knew who he was. The notion of just showing up anywhere in the world and getting what is rightfully yours without any papers, cards, fuss, or whatever has always intrigued me. That will be the reality for all of us very soon, whether we like it or not. Wherever we go, we’ll be recognized and known. Our reputation will follow us. This will be true for citizens, criminals, refugees, rebels, terrorists, intellectuals, journalists: literally everyone.

Future warfare

Guerrilla warfare and terrorism are based on the ability to hide and strike unexpectedly. That kind of thing is a lot harder when you are being tracked in such a way that you can’t escape it. So hard in fact that you would have to assume this style of warfare becomes essentially obsolete.

How would that work out in say Syria, Afghanistan or a similar future conflicts? It’s a rather large region with plenty of remote areas to hide in. How would surveillance work in such an area? This would require lots of networked equipment to be deployed. This equipment would not necessarily have to be very large, or expensive, or even durable. It would just have to be plentiful and phoning home. That sounds doable.

The reality in such a future war would be that e.g. the US would know the exact location of all individuals, including any enemies at all times. They would be able to track all their movements and observe behavioral patterns. They’d be able to identify these individuals as well through biometrics. And on top of that they’d be able to build up a historical archive of all this data. There would be very little left to the imagination. So, how would a terrorist, rebel, or anyone, hide in such a world when they’d be recognized and have their every move accounted for whenever they’d be in range of any kind of electronics? A single meeting with a known “bad guy” would literally mark you for life as a person of interest.

The simple answer is they wouldn’t be able to. There would be complete information on all their movements, including who they met, when, and where along with a complete log of all intercepted communication. Any strange disappearances, behavioral anomalies, etc. would stand out as well. In other words, such a hypothetical bad person wouldn’t stand a chance of plotting anything at all.

There wouldn’t even be a need to kill the individuals involved since it is hard to pose a serious threat when you are being watched.This last point is actually the most interesting since not only would it be unnecessary, it would be extremely hard to do in secret and get away with. The notion of doing things off the book gets a lot harder when everyone is watching.

Watching the watchers

What I’ve portrayed so far is depressing and scary. We’ll be watched at all times and lose the ability to keep secrets. That sounds pretty Orwellian. The naive assumption so far is that surveillance will be done by our governments and that those will continue to work the same way as they always have. This is the premise for all dystopian outcomes: a single, all powerful evil entity monitors our every move and acts in secret to control and dominate us and controls all access to resources needed to do so. While that makes for interesting movies and literature, the real world is a bit more complicated.

Surveillance, the technology needed for it, and access to surveillance data will be commodities. There is no good reason to assume that this would be restricted to a select few governments or countries that end up hoarding all forms of technology. In fact, many tech companies already have a better insight into who we are than most intelligence agencies. This raises some interesting issues about exactly who will be watching who and how they will make use of the resulting insights. It is unlikely that all these parties will be friendly with each other. Surveillance will be a game that involves basically everyone and one that won’t respect borders, laws, or common notions of decency.

Simply put, ubiquitous surveillance means that the watchers are being watched as well. All the time. This is an aspect of ubiquitous surveillance that has far reaching consequences because it makes acting against others in secret extremely hard, even for powerful states or people. Whatever you are going to do, you’d better be ready for that to become public knowledge.

Most of today’s more shady activities by governments, criminals, terrorists, cheating husbands/wives, wannabe presidential candidates, etc. are happening in secret away from public scrutiny and under the assumption that things can be covered up when needed. Secrecy relies on not being watched or exposed. What happens when secrecy becomes sufficiently hard that you can never be 100% sure that something is actually secret no matter if you are the president of a large nation with a huge army, or a farmer in rural Afghanistan? People will be watching each other on an unprecedented and colossal scale and the act of doing just about anything in secret will become hard and risky. People will know they are being watched, whether they like it or not.

This changes everything.

Auditable surveillance data

Given that putting somebody under surveillance has a high risk of being detected by someone else, people will want some legal protection against abuse of surveillance technology. The massive availability of surveillance data means it will be used in court cases as a form of evidence (including cases of abuse of surveillance data) by both sides. This will inevitably raise the question of how trustworthy that data actually is, where it came from, who handled it, and whether it was tampered with, and whether it was accessed lawfully.

Given the current level of productions in Hollywood, it is quite easy to imagine that video footage will be extremely hard to audit simply by looking at it. It might have been edited, modified, or be completely computer generated. How would you be able to tell the difference? As technology for surveillance improves, so does our ability to forge and tamper with that data.

What’s needed in a court, or indeed any kind of place where the stakes are high enough, is trustworthy evidence. This requires that there is an audit trail and this in turn can be facilitated by technology. For example, it is possible to digitally sign content as it is recorded, to log transactions involving the data in a bitcoin style ledger, and to use checksums and fingerprints to ensure there is an audit trail for any bit of information that captures its origins, alterations, and that allows for verifying that the data was not tampered with. Building systems that can do this is fundamentally possible with technology that exists today.

Not only is this possible but it is going to be essential to be able to make sense of surveillance data. It is going to be a hard requirement to be able to guarantee data was not forged, not tampered with, and to have rock solid audit trails on who accessed it. If you are acting based on information learned from surveillance data, it would be rather inconvenient to find out later that your sensors were hacked and that you’ve been watching an elaborately composed fantasy while being recorded doing so by everybody else. There is going to be some very strong incentive for people to make sure they are not being lied to.

Auditability will become a key enabler for legal protections. Just like it is a crime to open a letter (which is technically trivial), it will be a crime to access surveillance data without proper authorization or cause. And it will be possible to argue your case in a court with more auditable surveillance data. This is the loophole that will allow us to dodge the more dystopian outcome where the thought police monitors your behavior and preemptively singles you out for punishment (the classic Orwellian nightmare). Instead of legislating the recording of data, it’s the use of data that needs to be legislated and regulated.

Short term things will be messy

Given the above you might wonder why wait another 67 years? The challenges are significant and my impression is that things will get a lot worse before things get better. There are a few fundamental challenges: people are mostly not aware of the technicalities and especially politicians have proven to be highly gullible when it comes to using technology (e.g. the Clinton emails) or assuming friends always behave appropriately (e.g. the CIA putting Merkel under surveillance). Technology is simply not that far yet and kind of expensive to use.

There haven’t been any really big incidents yet with surveillance technology yet even though there have been a few embarrassing situations. A few states are definitely showing signs of Orwellian abuse of power; e.g. by insisting on compromised encryption schemes to make it easier to keep an eye on citizens.

But fundamentally, use of surveillance technology is still sporadic, naive and clumsy. It is still possible to evade surveillance if you know what you are doing. Surveillance remains far from ubiquitous, easy, or commoditized. Information is highly fragmented between intelligence agencies, companies and countries. This makes it possible for people to game the system. As surveillance ramps up, gaming the system will get much harder.

The Middle East, surveillance ground zero?

A possible ground zero for ubiquitous surveillance might be the middle east. There are lots of people in sparsely populated terrain and they seem to be hard to find or track. Recent wars there involve large but ineffective armies fighting people that are using guerrilla tactics. ISIL combatants are moving around large territory where it is easy to hide. They are performing guerrilla attacks with small units that hide amongst civilians. And they make use of the collapse of the local despotic regimes to cross borders, move people around, and operate outside the system. Traditional military power is useless unless you can pin point it at the right places and the price for getting it wrong is a blood bath that involves innocent bystanders, full coverage on youtube, twitter, and major news channels, and retaliation attacks on home ground by sleeper cells.

That, in a nutshell, is what ubiquitous surveillance solves. Imagine the ongoing battle in Mosul with perfect intelligence of who is hiding where, who they are, and their histories from birth to now covered in excruciating detail. It would be over in hours or minutes even. The whole point of that battle is that that intelligence is currently unavailable, which is why everybody is so reluctant to get involved and which is why everything has escalated so much. Bluntly put: that’s a fixable problem.

Since 9/11 intelligence agencies have been ramping up their capability in this area domestically and internationally but wars are still fought the old way. The obvious next step in this space is to specifically put war zones under surveillance with the goal of having perfect intelligence on who is active in that area, regardless of their status. This is where the proverbial gloves will come off because in a war, you don’t want to hold back. My guess is that this will play out over the next decade. It will be ugly and it will involve multiple nation states that are working with or against each other each with a deep mistrust of the other parties. Surveillance technology will be at least as disruptive as the invention of gunpowder was for warfare. Ironically, this might actually be the best shot this region has at some notion of a lasting peace. But what happens once that cat is out of the bag?

2084: After ground zero

The question is what happens after that. How will legal systems, international diplomacy and treaties evolve when the dust settles? This will inevitably be a slow process. Some countries will get it wrong, some other countries may not care, and some functioning democracies might actually do the right thing only to be toppled by less considerate non democracies. In the end, some system will emerge that puts some checks and balances in place.

Surveillance is a weapon that has both offensive and defensive capabilities. It can be used for good and for evil. Either we wipe each other out or we find a way to deal with it. With nukes, we’ve so far managed to not create an apocalypse; though we’ve gotten close a couple of times. I’m hopeful we won’t do it with surveillance technology either.