What will be the eventual future of privacy? How will privacy work? What will it cost? I’ve already started a thread on privacy:
My argument for why we desire privacy is that we want to prevent others from using certain information to harm us. Now we don’t know how hard it will be to keep up privacy in a future that is increasingly saturated with technology. It would seem that, as our world becomes increasingly hyperconnected and more and more sensors are deployed in the environment, our homes, and even our bodies, maintaining a high degree of privacy will become an increasingly hard prospect.
The Leviathan sees everything
Can that outlook somehow be avoided? What if we prohibit spying on others in general? Well, as sensors, spy drones, and their like become ever smaller, smarter, and cheaper, it would become increasingly difficult that a particular person spied on someone else. In 30 years, what stops the average Joe from printing out thousands of robotic spy flies that gather information on everything in sight? Either we spy on Joe ourselves, or we don’t know who releases all those spy flies that probably won’t even noticed. This should give us a clue: A privileged organisation, a secret service with the authority to spy on people, but with protection from being spied upon, could be used to prevent unauthorized breaches of privacy. In the end, we’d have to trust some kind of “privacy Leviathan” to protect our privacy from everyone else. But who protects us from that Leviathan? This is a troublesome question.
We could try to control that Leviathan somehow, but such efforts could threaten out privacy again. If we see what the Leviathan sees, and the Leviathan sees everything, then we see everything, and privacy is lost again. We could try establishing a network of organisations that keep each other in check, so that they don’t exploit their information on us in any due way, and we would like to have some democratic or economic leverage on those organisations. That’s probably the best we can hope for, under usual conditions.
Unusual conditions would be that the Leviathan was completely trustworthy, for some interesting reason. Perhaps it’s a kind of friendly AI that is incapable of betraying us. Or perhaps it’s provably aligned with our interests. In that case, we’d have what I call a “perfect Leviathan”. Otherwise, we’d have to deal with imperfect Leviathans with checks and balances.
Strategies of privacy fanatics
Our course, privacy fanatics wouldn’t want even a Leviathan to have access to their private information. But without a Leviathan protection against intrusion of privacy committed by third parties will become increasingly difficult and costly. Yes, you can protect your own home with lasers that shoot down spy drones. Even if there are nanobots that spy on you, you could lace your home with a nanobot immune system that kills off intruders. The walls of your house may be listened on with lasers that pick up the vibrations that are caused by voices inside, but you might want to opt for encrypted tight beam laser communication that carries signals for a telepathic local area network anyway. Outside of your home, maintaining your privacy will be almost impossible. You might hide your identity by smearing your face with paint that throws off face recognition, and wear stilts that makes gait analysis harder, but no normal human would recognize you anyway, and the chances that AI recognize you by your skeletal features is so high that no real privacy will be possible outside of your home. And of course the entry to your home is likely under constant supervision, so you won’t be able to hide when you are together with certain people in your home.
With the arrival of the Internet of Things, you would have to opt for “dumb stuff” that’s not connected to the internet all the time. If you have devices with cameras in your home, you might want to cover them in case that someone has access to them (which isn’t terribly unlikely). If you don’t resort to computers that don’t have access to the internet, at the very least you use Tor all the time to remain anonymous online.
You don’t tell anyone anything about yourself that isn’t already publicly known – unless you can trust those people not to propagate that information.
Yeah, hard times for privacy enthusiasts are upon us. Only the wealthy and paranoid will be able to maintain a relatively high level of privacy.
Everyone knows almost everything
So, why not get rid of privacy entirely? That way, at least nobody has an unfair advantage by knowing more about you than you know about them. Well, sure, that’s an option, but we are discussing the possibilities of privacy here, so this option only deserves a brief mention.
How does privacy work after all?
Ok, so privacy consists in the idea that you can prevent person X to know information J about you. You obviously know information J, otherwise you wouldn’t deem it worth protecting. There are certain strategies for maintaining your privacy:
- Spread containment strategy: the basic strategy for privacy is preventing others from gaining access to J. That’s by creating obstacles to the spread of information J. Non disclosure agreements and digital rights management are examples for such obstacles.
- Disinformation strategy: You make X unsure about the truth of J by spreading conflicting (untrue) information K.
- Access denial strategy: You prevent X from accessing information J.
- Prevention strategy: You prevent X from using J to your disadvantage.
- Distraction strategy: You make X focus on something else, and thus not think about J.
Let’s take a look at the different strategies in turn:
#1 Spread containment strategy
Ideally, information J is a secret that only you know about. In that case, everything is fine. Unless you need to let others know about J in order to make full use of that information – for example, if it’s some kind of invention that you need help for implementing it properly. Once you let others know J, it becomes a shared secret. The difficulty lies in preventing a further spread of J to people you don’t trust.
One method for dealing with that problem is by punishing those who share J with unauthorized people. That’s basically what NDAs can help with in the worst case. A more extreme variant of this is suing a person for treason for spreading the state secret J. In a very wide sense, copyright belongs to this class of strategies, too. An information J that is protected by a copyright is intended to be sold for profit, so uncontrolled spread of J would perhaps diminish profits (though that’s not necessarily the case).
A complementary strategy to copyrights is digital rights management that makes it harder to copy J freely. This strategy isn’t based on punishments (at least not directly), but on creating additional barriers that have to be overcome before one is able to copy J to other people. In reality, DRM basically makes it more time consuming to copy J, but usually that’s good enough.
#2 Disinformation strategy
This strategy works, even if J has spread to people who aren’t supposed to know about J. So, the strategy is that you invent some (usually false) information K that contradicts J and spread that to as many people as possible. Then you make go on to discredit those who still believe J to be true als gullible fools. The world may be spread between those who believe J and those who believe K – divide and conquer. The situation has stopped being about knowledge, but has become about beliefs.
Disinformation is more about damage control than about preventing harm in the first place, though it’s conceivable to use this strategy in a preventative way. It would even make a lot of sense to spread out false information K before J even has a chance to get leaked. By priming the people on K, any alternative J may look like a ridiculous rumor. Repeat a lie often enough, and it becomes a widely accepted fact.
#3 Access denial strategy
Now, let’s come to some seriously advanced and futuristic strategies. Let’s assume that X has access to J, and know that it’s the truth. With a sufficient ease of spreading information, and access to a comprehensive sensor network, this will sooner or later be a nearly inescapable situation. You would like to remove J from X’s mind, but that won’t help for long, since J is rather ubiquitous by now. What could be done, however, is denying X access to the information J, even though X has it in her mind.
How does that work? You insert a malware I term a denial veil into X’s mind. The effect of the denial veil is that X will always deny the truth of J whenever J pops up in her mind. Denial of the truth of J becomes of overriding priority. In every situation, X will find a way to rationalize why J must be false. There’s no alternative – because the denial veil doesn’t allow the existence on an alternative. The falseness of J becomes self-evident to X. And of course that’s not the case, because X suffers from a denial veil, but because J is obviously false. Everyone else must be a crazy conspiracy theorist, and completely deluded. For most practical purposes, denial veils are equivalent to deeply held convictions, so they aren’t terribly easy to detect – unless one is suspicious of all deeply held convictions.
But why not stick to a seemingly simpler approach and try to insert a deletion veil into X, that deletes J every time it enters the mind of X? This becomes problematic when X is faced with strong evidence in favor of J. Every time the deletion veil becomes active, X spaces out for a moment. The more X is confronted with evidence for J, the more X spaces out. At first, this may appear as an increasing pattern of concentration problems, but later on, it can become increasingly paralyzing. Denial veils avoid this problem, since they don’t cause such apparent neurological impairments, but merely strong convictions.
#4 Prevention strategy
An even more advanced concept than a denial veil is a conditional inhibitor. A conditional inhibitor inhibits X from using J as justification to harm you. This is a bit complicated. What the conditional inhibitor needs to do is to compare two scenarios. First, it needs to predict the consequences of X knowing J during a specific action of X. Then, the conditional inhibitor predicts the consequences of X not knowing J during that action of X. If the results of the first scenario are worse for you than those of the latter, the conditional inhibitor inhibits X from proceeding with the action in question. The actions that remain for X are those who are neutral with regards to the knowledge or truth of J, or those that actually lead to a better result for you, given the knowledge of J.
Conditional inhibitors can even work for information that isn’t very secret, after all. Assume for a moment that you could insert conditional inhibitors into X that prevent X from discriminating against you for being a woman, black, a criminal, or completely incompetent. That’s pretty crazy stuff. What’s scary about conditional inhibitors is that they seem to be more benign than veils, so the temptation to actually use them on people becomes stronger. Governments may feel tempted to prescribe the use of certain “sensible” conditional inhibitors on the whole population.
#5: Distraction strategy
Ok, so this last strategy can be used with low tech tools, as well as high tech tools. Let’s start with the low tech approach. So, X has finally found out about J, but, so what, you tell X about L. Woah! L, yes, L is much more important than L. Yes, why care about J, if L is so fascinating, and urgent, and captivating? L, think about L! Everybody is talking about L! We must react to L, no time for anything else? What? Has there ever been something else? No, of course not, we’ve all been preoccupied by L since the dawn of time!
And once distraction with L stops working, you throw M into the mix. And then N, and so on…
In the future we don’t even need to come up with things like L, M, or N. We simply insert a distraction veil of J into X. What the distraction veil does, is simple. Once X thinks about J, the distraction veil redirects the attention of X towards anything else that happens to occupy the mind of X.
In many ways, distraction veils are the perfect tool for keeping J secret in plain sight. Even if everyone knows about J, if everyone also is under a distraction veil on J, then it simply doesn’t matter. J becomes a perfect taboo. It’s not that people don’t know about J. Of course, people know about J, but nobody is able to think about J, or care about J.
A great demonstration of this idea of a distraction veil is found in the books of the “Night Watch” book series bei Sergei Lukjanenko. In the contemporary fantasy setting of the book series, there are magicians and other magical creatures, walking among us. But normal people are under a universal spell that makes them not care about all the strange things happening around them. They may see magical events, but are magically forced to ignore them, and afterwards they are quickly forgotten.
The technical implementation of veils and inhibitors
In our glorious cybernized future, the veils and inhibitors are artificial intelligences that interface with our minds. Given a scenario, in which it will have become very easy for nearly everyone to observe everything and everyone, they are the remaining tools which can maintain privacy (or collective ignorance, which is basically the same).
As with all technologies that affect the mind directly, the potential for abuse is incredibly high. Using those tools wisely will be a huge challenge for humanity, and its potential successors.