Censorship is surveillance, and privacy is a public health problem

My latest Guardian column is "Censorship is inseparable from surveillance," which discusses the fact that network censorship entails surveillance, and how this exacerbates the public health problem caused by our difficulty in evaluating privacy trade-offs.

There was a time when you could censor without spying. When Britain banned the publication of James Joyce's Ulysses in the 1920s and 1930s, the ban took the form on a prohibition on the sale of copies of the books. Theoretically, this entailed opening some imported parcels, and it certainly imposed a constraint on publishers and booksellers. It was undoubtedly awful. But we've got it worse today.

Jump forward 80 years. Imagine that you want to ban www.jamesjoycesulysses.com due to a copyright claim from the Joyce estate. Thanks to the Digital Economy Act and the provision it makes for a national British copyright firewall, we're headed for a system where entertainment companies can specify URLs that have "infringing" websites, and a national censorwall will block everyone in the country from visiting those sites.

In order to stop you from visiting www.jamesjoycesulysses.com, the national censorwall must intercept all your outgoing internet requests and examine them to determine whether they are for the banned website. That's the difference between the old days of censorship and our new digital censorship world. Today, censorship is inseparable from surveillance.

Censorship is inseparable from surveillance


  1. Censorship has always been inseparable from surveillance, you can’t censor something without knowing what it is, and where it’s occurring.  The methods of disseminating information have just gotten better, requiring more intrusive methods to catch them.

    Taking another look at your Ulysses example, people could have also smuggled in copies in their luggage as they came and went from the UK.  To truly censor this book would have required customs to go through people’s luggage, etc.

  2. I can’t agree with this definition of “public health problem.”  A “public health problem” should be something where your actions affect the rest of us– a prime obvious example being communicable diseases (and refusal to get vaccinated against it.)  Obesity and smoking are “public health problem” only to the extent that they affect the rest of us (and yes, they do.)

    In fact, I’d say that your recommended policies of “instituting a minimum drinking age, prohibiting public smoking and so on – and conduct public education campaigns” are a *lot* more like censorship and surveillance than obesity and smoking themselves.

    Your comments about the necessity of governments to prevent the “public health problems” of smoking and obesity– because what people want to do doesn’t fit with your conception of their long-term best interests– is exactly the same logic that led to alcohol prohibition and, yes, prohibition of obscene and pornographic material.

    Censorship and surveillance isn’t better when it’s *your* kind of censorship and surveillance.  Even if you’re really sure that people *should* want to avoid being obese, and that their pleasures of smoking can’t possibly make up for the risk of lung cancer.  You shouldn’t make that decision, just as people shouldn’t have the right to say that reading the wrong sort of literature will make you unhappy or give you bad thoughts, and engaging in too much pornography will destroy your relationships, so let’s discourage and censor them.

    1. Good point. Calling privacy the equivalent of a public health problem is nonsense. Public health is issues are all about externalities. But privacy involves no such externalities.

  3. The Comstock laws empowered inspectors to censor US mail– and customs officials have always rifled through luggage, so this is nothing new. Perhaps the new regimes rely less on discretion, and are somehow less humane. 
    The new Tyranny will consist of unjust laws, perfectly applied, without the inconvenience of common sense. 

  4. “by instituting a minimum drinking age, prohibiting public smoking and so on – and conduct public education campaigns to help people appreciate the potential future effects of their present-day causes.”

    In other words, Cory believes that good censorship, surveillance, and influencing people’s moral choices is good, but bad censorship is bad.  Luckily, he’s just the person to decide what the rest of us should be doing.

    Is it obvious that in the long run people will regret overeating or smoking, but that they won’t regret viewing some desensitizing violence or pornography?  Or reading some extremely misleading lying tracts (say, Mein Kampf or the Protocols of the Elders of Zion?)

    Cory isn’t against censorship.  He just wants censorship that makes public health problems “better” instead of those that “make the problem worse.”  And he sees nothing wrong with imposing his own values.

    1.  I can’t agree here. You can argue plausibly that, say, prohibiting public smoking is wrong, intrusive, a government overreach, and requires surveillance.  But it’s ridiculous it to describe it as censorship; it makes the word so broad as to have no meaning. (Unlike the required health warning on cigarette ads, which clearly is censorship of one form.)

      Likewise, Cory has never claimed that “influencing people’s moral choices” is inherently bad; we all do it all the time.  He just thinks we should do it by giving people more information, not less.  You can agree or disagree with Cory on this point, but he’s consistently anti-censorship.

      1. Does a “transvaginal ultrasound” count as “influencing people’s moral choices by giving them more information?”
        I’m not so sure that moral choices are really up for grabs here. Is smoking morally objectionable, or is it physically harmful?

  5. Yes, evaluating trade-offs between features isn’t easy. That includes privacy features.
    But think about how the industry solves this problem in general. Reviews is a common solution (and there are others), which doesn’t require forceful regulation. 

    Before you use an App or website, check out the reviews and reputation from a trusted privacy-sensitive group, such as EPIC or Boing Boing. There could also be reviews in Consumer Reports and Zagat website ratings. Browsers could simplify such interactions.

    There is nothing that unique about privacy. It is simply another feature in a world of complex products (cars, technology, …). The best way to protect consumer is let competition figure out the set of features that can economically be produced and offer maximum consumer satisfaction.

    1. Funny you should use cars as an example for that, when consumers are protected largely by safety requirements. Back in the day, competition didn’t even ensure you would get seat belts.

      1. Regarding seat belt legislations, see wikipedia:

        “The effects of seat belt laws are disputed by those who observe that their passage did not reduce road fatalities. There was also concern that instead of legislating for a general protection standard for vehicle occupants, laws that required a particular technical approach would rapidly become dated as motor manufacturers would tool up for a particular standard which could not easily be changed. For example, in 1969 there were competing designs for lap and 3-point seat belts, rapidly-tilting seats, and air bags being developed. But as countries started to mandate seat belt restraints the global auto industry invested in the tooling and standardized exclusively on seat belts, and ignored other restraint designs such as air bags for several decades.”

        I suspect the industry’s hesitation to advertise seatbelts reflects consumer preferences at the time, which also explains why people tended not to wear seatbelts even when they were available. Nowadays, people’s preferences have changed and most wear seatbelts regardless of enforcement.
        Also, I suggest you look at the air bag legislation, which is far from a slam dunk (air bags increase mortality and morbidity at some speeds).
        Finally, I would suggest that some important safety features are way ahead of legislation (self-driving car, OnStar calling emergency responders, …). And further you should be aware that some brands do emphasize and advertise safety above other features.

        1. The effects of seat belt laws are disputed by those who observe that their passage did not reduce road fatalities.

          Weasel words with one citation at the end of the paragraph that mentions, but doesn’t link to, a newspaper article from 1969.

          1. Seat belts do have an effect (they save your life in accident), but that is not the only effect (people drive more recklessly with them). Seat belt legislation and safe driving habits and norms are simply not the same thing.
            If you look at historical data, you will find a stable 30K deaths/year in the US.

            Again, my main point is not whether seat belts are a good trade-off, but that at the time there was not much consumer demand for them. Legislation made them mandatory in cars, and yet it took a long time for people to wear them. Seat belts are not a counter-example to how competition provides features people want.

  6. All this censorship talk and legal maneuvering by governments is the last gasp of a failing business model.  There is absolutely no way to put the internet genie back in the bottle.
    Workarounds to these laws will appear before the ink has time to dry.  That’s the reality.

  7. I want to go farther and say that censorship is an expression of corporate interest. To censor something requires authority, and to have authority one must have economic power to control what is censored. In the case of the U.S., with these assumptions, I feel it is not provocative to say that government censorship is an expression fo corporate greed.

    That is, any effort the U.S. makes to regulate information it does so as direct expression of its status as a corporate entity. In order to censor something, one must be a participant.The question should be _What kind of entity wants to censor?_ The government either censors as a Corporate Entity or an Authority. There are difference consequences given the status of the entity in question. The Problem is that governments (China, India, etc) have tried to censor as Authorities, while the U.S. is attempting to censor as a Corporate Entity (defending “economic creativity” because establishes economic substrates and monetary flow structures for the country in which the government exists: think of it this way, if “We the people” makes sense, then the government is the largest corporate interest possible; all people are the government’s “Employees”: “We the Employee”).This is why Google has to be “the smartest in the room”. In the Joyce example, we have a government functioning as an Authority. This is what usually happens (say, with fascist governments, reactionary religious institutions, etc). It’s worrisome when the government uses the tactics and schemas of Authoritarian systems in order to defend their Corporate Interests. The Problem is that our government is try to self-preserve, but it is currently unsure of its own scope. This is why Federalism was so important. Unlike the Internet, where we have a logical foundation to determine all sorts of answers to all sorts of problems (even possibly NP Hard Problems!) Governments of brick and stone, of centuries ago, simply couldn’t imagine the range of logistical problems that we face today. With large government, the onus of solving those problems doesn’t change (imagine a world where Google is our government), but yet the government is too “small” to handle the scope of its material assets and related parties.So we see now that the U.S. government is trying to dance between defending us and defending itself.

  8. re: “When Britain banned the publication of James Joyce’s Ulysses…”

    In this one case – they were doing everyone a favor.

  9. “Today, censorship  is inseparable from surveillance” is one of those rare, BIG insights and memes that can change a lot of minds. Mr. Doctorow, you have outdone yourself – I think this article will be very influential, whether you realized it when you wrote it or not. Kudos!

    Would love to hear more thoughts on the relationship between surveillance and censorship, (in both directions). Worthy of several follow ups IMHO.

  10. So … 15 comments in and nobody is pointing out that the underlying technical assumption by the OP is incorrect ?

    Yes, you could conceivably set up a single, personalized filtering rule for every single client on the network.  But instead of 10 million individual rules, you can just put a single rule on the server side:

    deny ip from any to me [1]

    So you can block all traffic without identifying, or personalizing, the censorship at all.  Blocking (and “naming”) individual clients is totally unnecessary.

    So in this example, while censorship *might* imply surveillance, it doesn’t need to, and probably wouldn’t in any real world scenario.

    [1] Not pseudocode – that is actual ipfw syntax.

  11. >Ironically, from a human rights perspective, censorship with surveillance is better than censorship on its own. 

    But you just said that censorship without surveilance was impossible. the rest of the paragraph/colunm reads like it should be:

    >Ironically, from a human rights perspective, surveillance with censorship is better than surveillance on its own.

    which makes a lot more sense.

Comments are closed.