The 22nd century C.E. introduces a new concept of "smart city" in which everything is automated and requires little or no human intervention. For example, the daily household chores such as washing clothes and preparing breakfast will be taken care of by the smart kitchen. This is thanks to a breakthrough in quantum computing and heuristic algorithms which led to the development of a very powerful, central artificial intelligence. There are thousands of microscopic cameras throughout each house that provide live feed to the A.I. for processing and analysis.

In order to protect the safety and interest of all its citizens, the A.I. can dispatch resources to areas without delays; response time is greatly reduced. Of course, there will be actual humans to help maintain the system and ensure all is kept in order. The government passed a law to legally gain all rights of access to every live feed in the city. So far, crime rates are kept at an all-time low, and there are no new reports of unnatural death. This system can even suggest what's on your menu for lunch and monitor your diet, and it will send for the closest doctor should the situation becomes dire. The system thoroughly scans your entire body, including micro-expressions, and compares them with past records, to accurately predict what's on your mind every minute of your life.

Despite all the conveniences provided by this system, some find this intrusive. By the way, the government understands that there will always be opposition and that the number of the discontented must be kept to a minimum. Knowing that tampering with the cameras is against the law, how can privacy possibly exist?

  • 1
    $\begingroup$ This time is now. Or soon. insecam.org - here is your access to a lot of cameras. Just plug an AI ;) $\endgroup$
    – Mołot
    Commented Dec 6, 2016 at 7:17
  • 2
    $\begingroup$ 1. As far as we currently know there is no such thing as a "microscopic" camera; optics does not work that way. 2. Privacy for the masses is a recent (19th century) invention; maybe its time in this world will not be very long. 3. See Great North Road by Peter F. Hamilton for a great fictional treatment of a police investigation in such a future. $\endgroup$
    – AlexP
    Commented Dec 6, 2016 at 8:02
  • 4
    $\begingroup$ 1984 was a warning, not a training manual! :) $\endgroup$
    – Ghotir
    Commented Dec 6, 2016 at 15:26
  • 1
    $\begingroup$ What definition of privacy do you want to use? A society with a singular individual (the AI) with that much power will have redefined the word to mean something different than we mean it today. This is similar to how privacy has changed meaning with advent of the internet (and social media). Also, how did the law preventing tampering with the cameras get passed. The details of that event will dramatically shape how such "privacy" issues are addressed. $\endgroup$
    – Cort Ammon
    Commented Dec 6, 2016 at 16:03
  • $\begingroup$ Logistics, optical physics, and cost aside, I think the real question is how a society would ever agree to this. Interesting idea though. $\endgroup$
    – Zxyrra
    Commented Dec 7, 2016 at 4:50

10 Answers 10


Big Brother is watching you

You've said it yourself. Privacy doesn't exist, there are cameras everywhere and the government can access them all. Along with that you've got an AI watching them, that was the weakness in the 1984 Big Brother system, which was much more personal.

The only way to get away from the endless cameras in the city, is to leave the city. Go and live in the Scottish Highlands, find a place with no mains power and no mobile signal. There you will find privacy.

Except you won't, because every other person who knows you exist will know your business, they'll talk about the comings and goings of the new person who's taken the hut on the hill. Your life will be the talk of the village.

Privacy is really found in the anonymity of the city, a place where everyone can see you and nobody cares.

  • 1
    $\begingroup$ Or take an EMP backpack and detonate it whenever you need some personal time... $\endgroup$ Commented Dec 6, 2016 at 17:30

So you have cameras everywhere, covering 100% of house/city limits.
Scanners that monitor your health and expressions 24/7.

The only way to have privacy (in any form), in a city like yours, and still live there, is to be anonymous.

Anonymity is obtained by staying out of 'alert" situations, so no breaking the law, monitor your own diet/health so as to reduce health issues, obey deadlines on taxes, etc

Do this for a long time and nobody will waste time on someone as boring as you.
Do this for a long time and become the perfect robot-citizen or something that resembles the Equilibrium film.

On the other hand, there will be a percentage of the population that values privacy a lot. These people will have to engineer ways to "hide" their signal from your A.I. So, small devices that give false signals, or cause time-dilation on camera recordings or anything similar.


The fundamental flaw in the situation you are describing is the Local Knowledge problem, as described by F.A Hayek.

Since ephemeral and fleeting information is distributed throughout society, gathering it to a central location for processing, analysis and distribution of orders based on that knowledge takes finite amounts of time, and large amounts of bandwidth. This is true regardless if Big Brother is at the top to the pyramid or HAL 9000, the only real changes are in terms of scale and speed.

This sets up a situation where negative feedback loops are induced. The information takes finite amounts of time to move up, be processed and acted upon, then more observations take place to determine the efficiency of the action. Each cycle is based on outdated information, and so each action is cumulatively more incorrect, and the incorrect action induces more changes to the system, which need greater action to overcome and so on. The chaos and eventual collapse of Socialistic economies ranging from the USSR to Venezuela are all real life examples of centralized "command" economies falling apart due to the local knowledge problem, and mixed economies generally can be graded on the amount of centralization based on PPP, GDP and other economic comparisons between them; in general, more centralized states do worse than more decentralized ones.

The military expression of this idea is the OODA loop (Observe, Orientate, Decide, Act).

enter image description here

Conventional militaries can attack each other by running their OODA loops faster than the opposition ("getting inside the OODA loop"), so the enemy commander is observing and orienting on things which are no longer relevant by the time he makes a decision and acts upon it. In this construct, while it is unlikely a human can get inside an AI OODA loop, it may be possible to make use of the overall slow loop (due to the massive amount of data being processed) by inducing errors in the observations (ranging from spoofing cameras with video loops, IR lamps to cause flares in the camera to arranging for flashmob to do large scale, if relatively innocent things like performing show tunes, to distract the attention of the observer. Breaks in the Observation translate into errors in the rest of the cycle.

Since the scale and scope of the resources available to the State are so much large than available to an individual or even a group, it is difficult to think of how to go beyond this small scale harassment, but if you are writing fiction, then you now have the entry point to start.


Are you assuming that there is one governement governing the entire planet?

If there are different countries with different governments, some poor or rebel countries would probably not adhere to this "monitoring" policy. You could just escape there.

  1. Cameras have blind spots. The main actor in 1984 had a blind spot in his flat, where he hid from the camera and started writing (if I remember correctly).

  2. The surveillance cameras are in the city. How much of the outskirts, countryside are covered? The opposition may meet farther away from the city in an old farm.

  3. Go underground, literally. Hide in the sewers. They stretch over miles and miles below any major city. Sewer accesses are at (1) blind spots of the camera.

  • 1
    $\begingroup$ Would such an intelligent AI really have blind spots at such strategic places? I don't think so. $\endgroup$
    – Christoph
    Commented Dec 6, 2016 at 10:38
  • $\begingroup$ Take a look at London, one cities with the highest camera density in the world. You'll find blind spots there. And in the fiction story, there may be other issues, why a camera system cannot watch every place permanently. There's always some blind spot somewhere, at least for a short amount of time. $\endgroup$ Commented Dec 6, 2016 at 11:17
  • $\begingroup$ @Christoph The issue there is that there is a feedback loop. Places become strategic because they are in a blindspot. $\endgroup$
    – Cort Ammon
    Commented Dec 6, 2016 at 19:13
  • $\begingroup$ @CortAmmon It seems you assume the cameras would be a) static b) placed only by human intelligence. There's certainly not such a simple way to escape an AI that is able to track millions of lives at once. The only way to hide would be to disable cameras in a way that cannot be distinguished from pure randomness. A simple statistical anomaly of people moving towards blind spots would be enough for such an AI. Such an AI would certainly also be supported by a vast variety of drones. Especially in the 22nd century. $\endgroup$
    – Christoph
    Commented Dec 7, 2016 at 9:46
  • 1
    $\begingroup$ @Christoph *grins* No, that's not quite the assumption I make. I make the assumption that there's a feedback loop where the AI tries to move the cameras to places of importance, while the people try to make places of importance where there are no cameras. It becomes a game like Chess or Go at that point, but with far more nuanced rules. Of course, if the AI is at the point where they literally have a perfect camera/microphone following every person everywhere at all times, then it is fair to say that privacy is in a bit of trouble because we have been outclassed. $\endgroup$
    – Cort Ammon
    Commented Dec 7, 2016 at 14:51

This is a trope in science fiction.

John Varley came up with an answer in Steel Beach. And substantially the same answer was part of the plot in an episode of the TV show Elementary, "The Ballad of Lady Frances".

The monitoring of all that electronic snooping has to be done by computer. In Steel Beach it was the city central computer for the moon. In the Elementary episode, it was part of a system of microphones in New York City that listened for gunshots.

The principle of privacy enters like so. The computer listens for everything. But it keeps it all secret in encrypted archives. Only when some key feature or aspect arises does the computer divulge any of the information to any human. The parameters for this are set such that privacy is maintained in cases where no reasonable suspicion of a crime exists.

So the microphones listening for gunshots record everything. The idea is the sensors can triangulate and determine where the shot came from. But the recordings are absolutely sealed and unavailable to any human unless a gunshot is detected. Then the recording for 1 minute before to 10 minutes after the sound of the gunshot is released to law enforcement. This way, if you happen to be standing beside a microphone on the street, the recording it makes won't be available to the cops unless the sound of a gunshot happens while you are there. And then only the minute before the shot.

So the central computer on the moon is watching and listening to everything. But only when certain parameters are exceeded does it divulge anything. For example, two people are having a confrontation in a bar. It gets out of hand and one of them shouts for the cops. This triggers the CC, and the CC then summons the police, and makes available to them the portion of the recording from the point the confrontation started.

No it's not perfect. The imperfection was a big source of the drama in those two stories. But then, the situation before technology was not perfect either. You could easily have somebody peering in your window or pressing a glass against your bedroom wall. Demanding perfection isn't realistic. What we can demand is a situation that is good enough that we can tolerate it.


If you want to speak in private, go to the dance floor of a busy and noisy night club. Unless someone bugs the ear of the person you're talking to, they'll have no way of hearing what you're saying.

For other private meetings, hide in the crowd. Meet at a big event like a crowded concert or football match - obviously not where seats are pre-booked, but where everyone is just herded in to a tight location. The AI will most likely be only able to see your head and shoulders, so any suspicious transacting can go on below the waist.

  • $\begingroup$ I'm not sure how feasible the audio part really is. If we assume that each camera has at least one microphone (or even better, a beam-forming array of 2 or more, but that calshes with small size assumption), this approach would fail. Field researching that is called Bling Source Separation. We already have systems that can separate one voice from background static, that can unmix voice from music, unmix male from female etc. I believe that we can safely assume that BSS techniques will be even more advanced in future. $\endgroup$
    – AndrejaKo
    Commented Dec 6, 2016 at 14:17
  • $\begingroup$ What prevents lip reading in the OP's scenario? $\endgroup$
    – user
    Commented Dec 6, 2016 at 16:08
  • $\begingroup$ @MichaelKjörling ventriloquism! $\endgroup$
    – komodosp
    Commented Dec 6, 2016 at 16:11
  • $\begingroup$ @colmde You need to go watch 2001: A Space Odyssey if you think this will fool AIs $\endgroup$
    – Kys
    Commented Dec 6, 2016 at 18:40

At this point there is no privacy.. Except for your thoughts.

With current technology you can even scan human heartbet rate from a low-res webcam stream (see https://youtu.be/O3Bv0XhuwEM or https://youtu.be/H-mkmCf9jhA?t=2m19s), the AI can work as an automated "lightman group" from "lie to me" series

But there is a way to escape.

  1. First, everything you do should be counted as "normal", or should be filtered down as "noise".

Since all people are surely irrational, AI know this and will not be alerted for many actions you might find useful to trick AI.

  1. Second, you should find a person to communicate with privately.

If you want some privacy, just think, but there is little value if you can't share your thoughts.

  1. Third, you must use your language knowledge, humor, hints, grammatical and spelling errors

No matters how good AI is, even if it knows every possible hint ever, it will always be based on some sort of probabilistic analysis, which is always huge trade-off between false positives and false negatives.

Since safety of men should be ensured, eliminating false positives will have a huge impact on AI's ability to effectively understand humor, hints, etc. and will make a room for smart people to communicate securely.

Didn't you tried it with your friend's stupid little browser at home?

Also there is real-life examples: AFAIK, misspelling was used to effectively harden weak cryptography during wars. Try breaking even simple chiper (without a computer) where you can't even understand plaintext behind it.

Why AI? Well, because no amount of humans can effectively monitor those cameras after all. And even if they will try to - now "surveillance army" are first candidates to riot, not the citizens.

P.S. Russians already know this, and NATO just found out!

NATO analysts studied Russian comedy TV programs over the past nine years and concluded that they serve the Kremlin's political objectives and are directed against the West.

P.P.S. There is also many other things that "Easy on Humans, Hard on Bots".


I realize that this question is nearly three years old, however it just popped up on my feed for some reason, and I had a few thoughts that hadn't been mentioned yet.

Your anonymity lies in your ability to avoid being recognized by the AI. Therefore, people who want anonymity will:

  • Wear nondescript clothing and featureless masks which cover the full head. No skin will be visible. There might be a standard set of gray, baggy coveralls and pullover hood/mask which people wishing privacy will wear.

  • Avoid speaking in public or use voice distorters. Although not specifically mentioned, audio monitoring is implied in this scenario, and a voiceprint can also be definitive identification.

  • Live a cash-only (or even moneyless) lifestyle, as the use of any kind of credit can be tracked.


  • You don't say whether the cameras have thermal imaging or can otherwise see in the dark. If no, people can perhaps go to "privacy clubs." They walk into the club, walk into a dark room, change (or exchange clothes with another person), then anonymously walk out a different exit after a random amount of time. If they fear they're being tracked, they might change in the first privacy club and then go in/out of several more, exchanging clothes with other people along the way, to help elude pursuit. They can do this anyway, even if thermal imaging is part of the package, but they'll just need to do more to elude tracking.

  • If thermal imaging is part of the package, there may be thermal camouflage - i.e., vests or suits, which distort the wearer's thermal signature, perhaps making men having the thermal image of women, or vice-versa.

Several (if not all) of these methods were used in the book The Light of Other Days by Arthur C. Clarke and Stephen Baxter.

In all likelihood, you will have a percentage of the population who will choose to live their entire lives using anonymizing means, such as this.


Your scenario is flawed: a democratic government does not pass laws

The main problems with scenarios of this type is to assume that...

  1. the government passes laws
  2. the people do not care about privacy

Let us start with the second point. Authors that dream up this sort of scenario seem to think that they, and only they, but none of the people in their fictional world will ever hatch the thought "But what about privacy?", essentially reducing their whole population to mindless sheep.

Well that is the case then there is no problem because your fictional population do not care.

If they do care however, then that also solves the problem because then they will not let their government get any such wide- and far-reaching access to personal data.

This brings us to the first point. Do recall that the first thing that happens when you form a democracy is that you take legislative power away from the government and place it in the hands of the legislative assembly, i.e. parliament, congress, riksdag/Reichstag... whatever it is called in your country. Check the constitution of any democracy worth its salt and you will find very early on that it declares that power stems from the people, not from the government.

Example: the Swedish Instrument of Government

1 kap. Statsskickets grunder

1 § All offentlig makt i Sverige utgår från folket.


Chapter 1. The foundation of the constitution

Paragraph 1. All public power in Sweden stems from the people.

So the job of the government is not to decide over the people but to execute what the people — usually through their representatives — have decided. And the people usually are not very inclined to decide that the government should do things of that sort to them.

So when you say...

the government passed a law to legally gain all rights of access to every live feeds in the city

...you are essentially failing social science class because the government does not pass laws; the legislative assembly does.

If you want this sort of scenario then you have only two options:

  1. Your lovely city is not part of a democracy but some kind of totalitarian rule that does not have the kind of legislative assembly that decides according to the will of the people. In that case privacy is most likely the least of your worries and as such you have a large underground movement already. If not, then the people are probably quite content and accept this kind of rule. Problem solved.

  2. The people do not care about privacy. They are just fine with the authorities glimpsing what they do all the time and as such decided to grant them access to everything. And if so, this is the explicit will of the people. Problem solved.

Edit after critical comments

In the case of people that still wish for the possibility of privacy, there is the question of what provisos, exeptions and excemptions the legislative assembly provided when passing the law.

  • Did they really require that every inhabitable space must be monitored?

  • Did they really require that all personal data must be available at all times to all parties that have the least bit of curiosity in them?

If the answer is "Yes and yes, it is all or nothing", then I again say: this is not realistic, and the answer to the question is: you are out of luck, you cannot have privacy in the city. Because now you are in such a situation that even disconnecting yourself from the grid will flag you and direct attention to you. If you want privacy in this scenario you must leave the city, or evolve telepathy.

If you want to stick to the least bit of realism though, we should look at how privacy laws and personal data/integrity laws have evolved for the past 40+ years. There the trend is very clear that personal data is considered private and that you have the right to keep it that way, and only if the state can show they have due cause may they peek at that, which is then done under the process of checks and balances. And private actors must seek your permission to keep and process your data, and they may not cross-compile data with other actors.

So unless this trend is broken, there you have your privacy: the data is not universally available. We are already today considering your scenario as completely unacceptable from the start. Which means that when this law of yours is passed, there will be these kinds of safeguards.

Otherwise - as I said above - your only option is to leave the city or develop telepathy. And since the scenario is in the 22nd century CE, this is not unthinkable. Already today we have rudimentary brain-computer interfaces. Developing brain augmentation devices that allows you to communicate silenty, encrypted and obfuscated with others is next to an inevitability. Since you have not abolished Freedom Of Thought, there you may have your privacy.

  • $\begingroup$ The government may not be democratic, and even in a democracy all sorts of unpopular laws get lobbied into place. $\endgroup$
    – o.m.
    Commented Dec 6, 2016 at 16:40
  • 1
    $\begingroup$ @MichaelKarnerfors Respectfully, the bottom of the answer decomposes the situation into the two tautologically possible circumstances: either there is a resistance or there is not. OP did not ask for a critique of the scenario, nor did OP ask whether privacy should be a priority, nor whether a resistance exists. OP asked, and I quote, "how can privacy possibly exist?" Your response does not address that question. As I understand it, on StackExchange sites, it is expected that answers attempt to answer the question. $\endgroup$ Commented Dec 7, 2016 at 8:30
  • 1
    $\begingroup$ Let us continue this discussion in chat. $\endgroup$ Commented Dec 7, 2016 at 9:15
  • 1
    $\begingroup$ This would be great if your first statement weren't factually incorrect: a government, being a collection of people who govern, does pass laws. $\endgroup$
    – Frostfyre
    Commented Dec 7, 2016 at 13:11
  • 1
    $\begingroup$ I agree with Frostfyre that this is argumentative and not an answer. This is the kind of post we should avoid on this site to keep it more objective and less troll-ridden. Recommending deletion. $\endgroup$
    – kingledion
    Commented Dec 7, 2016 at 14:53

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .