Privacy rights and cognitive bias

Protection of privacy is a hot topic.  Hardly a day goes by without concerns over protection of privacy hitting the headlines with real impact (today’s example is “Google yields to privacy campaign” in setting their cookies to auto-delete.)  It seems clear that there is a general presumption in favour of privacy, in the sense that if something is seen to invade privacy this is a prima facie reason for stopping it, and the person wishing to go ahead bears the burden of justification.  But is this privacy presumption a rational response to the threat of invasive technology, or is it the result of a cognitive bias?

While I don’t work directly in the area, as a (somewhat bemused) observer I’ve always felt that there is a mismatch between the strength of feelings regarding privacy and the strength of the substantive arguments.  I think it is fair to say that in much of the debate as reported in the media no argument at all is made in favour of privacy.  It is just accepted as presumptively good.  This in itself suggests to me that there is a cognitive bias at play, even if there are ultimately good arguments for the privacy presumption. 

Moreover, while there are certainly good arguments in favour of protecting privacy in specific circumstances, I have never come across a sound policy argument that justifies a general presumption in favour of privacy of the breadth that seems to exist in the public discourse. One good policy argument in favour of privacy rights is procedural:  it helps prevent arbitrary enforcement of laws by police.  While this argument has considerable force in some contexts, it doesn’t explain the breadth of the privacy presumption:  Google puts cookies on everyone’s computer.  More generally, privacy concerns are just as prevalent in commercial and individual contexts where arbitrary law enforcement just isn’t at issue.  Another common position is the ‘bad laws’ argument – marijuana / surfing porn / speeding (pick a pecadillo) really isn’t that bad, and the presumption of privacy reduces enforcement of these bad laws.  The easy response to this is that the presumption of privacy also reduces enforcement of good laws, and most laws are good laws so the presumption must be a net negative with respect to law enforcement.  Of course, this isn’t an exhaustive analysis of the substantive arguments for privacy rights.  But the thinness of the arguments that are most often raised in favour of privacy inclines me to look for an explanation of the privacy presumption based on cognitive biases.            

I can think of two types of cognitive bias that might lead to the privacy presumption.  First is simply the salience bias.  We’ve all done something wrong that we would like to keep hidden (well, not me, but most people), and the consequences to us of being caught are far more vivid and easily imagined than are the indirect consequences of better enforcement against everyone else.  This is essentially the bad laws argument with a psychological twist.

A second possibility is that we might have evolved a direct cognitive bias in favour of privacy.  Suppose that free flow of information is in fact that the best social policy.  This would set up a classic prisoners’ dilemma:  the best case overall is if no one keeps information private, but the best case for me is that I keep my information private and everyone else reveals theirs.  Since everyone has the same reasoning, everyone elects to keep their information private, even though free flow of information would be substantively desirable.  Private information was no doubt as valuable in the ancestral environment (where is the bees’ hive, who am I having an affair with) as it is today, and a cognitive bias in favour of privacy might have evolved in response.

Of the two cognitive arguments I like the second one best, but that could just be because it’s more interesting than a plain old salience argument.  In any event, I find the cognitive arguments much more persuasive as an explanation of the general privacy presumption than are the substantive arguments.  Are there good general substantive arguments that have I missed?  Is there a cognitive bias in favour of privacy?  Are there other good cognitive arguments that would explain the privacy presumption?

GD Star Rating
loading...
Tagged as:
Trackback URL:
  • Stuart Armstrong

    Part of the bias may be because companies and governments are fond of intruding on privacy without warning. Instead of a clear exchange – you trade your privacy, under these conditions, for these benefits – the experience of most people is that their privacy has been abused and after that the abusing institution comes up with arguments why the abuse was fine.

    The ferocity with which people rail against privacy abuse seems all out of proportion with their willingness to sign up for loyalty cards and such, suggesting that it’s the “abuse of privacy” not the privacy itself, which is important.

  • http://profile.typekey.com/hollerith/ Richard Hollerith

    I for one agree with this blog entry that (genetically-transmitted, species-universal) cognitive bias is the explanation for most of the human desire for privacy — and other ways to avoid being held accountable.

  • Martin

    The presumption of privacy comes from the fact that, at least in the United States, it is a “right” derived from the Fourth Amendment. So you don’t have to have a persuasive argument to defend your privacy, just as you don’t have to convince anyone that you deserve to live if someone tries to kill you.

    A propos, you may find this interesting: http://tinyurl.com/yq92kx

  • http://profile.typekey.com/robinhanson/ Robin Hanson

    We each have a private interest in our own privacy, both because we want to control access to our secrets, and because privacy is a marker of status. We each have an interest in reducing the privacy of others, but perhaps we are not sufficiently aware of this interest?

  • josh

    “I think it is fair to say that in much of the debate as reported in the media no argument at all is made in favour of privacy. It is just accepted as presumptively good.”

    So is deliciousness.

  • Hollywood_Freaks

    I think one significant reason people are reluctant to give up privacy is the belief, whether accurate or not, that those who receive bits of information about other people will nearly always place it out of context. To stop this near-inevitable misrepresentation of themselves by others, people use privacy in order prevent an opportunity for other to misrepresent in the first place.

  • Stuart Armstrong

    Privacy now has a value – companies will make money from your information. It’s become economics, not morality.
    Anyone know if this has changed the debate much?

  • TGGP

    Although I am a libertarian, I find myself sort of the odd-man out when it comes to privacy. I don’t believe in any such right (not just in the sense that I’m an emotivist/Stirnerite rather than a natural law man), both in the sense that I consider the 4th amendment to be a protection of property and in that I don’t think a right to privacy is that desirable. I look forward to David Brin’s “Transparent Society”, even though I don’t think that is something libertarians qua libertarians ought to be in favor of (not that they should be against it!).

    most laws are good laws so the presumption must be a net negative with respect to law enforcement.
    My guess is that the majority of laws are bad laws, but the good laws tend to be the most important. How does the final balance come out? I don’t know.

    I found “I’ve Got Nothing to Hide” and Other Misunderstandings of Privacy to be one of the least persuasive things written by a professional I’ve read in a long time. I haven’t read the authors’ earlier papers in which he tries to define privacy, but in this one he basically seemed to say it can’t be defined, nor did he give much reason why privacy is desirable.

  • savagehenry

    I was going to mention the “I’ve Got Nothing to Hide” paper as well. It certainly made me think a great deal. I think the author did make the point that privacy is desirable because it benefits society as a whole. One of the ways it does this is that it ensures there won’t be a chilling effect on peoples activities. If the government (or anyone) knows everything you do odds are you’ll think twice before you do anything, even something as simple as talking with a friend or making a purchase.

    I think Norman is most definitely right though, even though there are some good arguments for privacy rights (even if they aren’t clear cut and easy to understand or define), most people seem to take a pro-privacy stance simply because they feel it’s right and anyone who comes up with reasons against privacy is obviously a bad person.

  • thuddmonkey

    I think the “bad laws” substantive argument may be more powerful than the post gives it credit for. I think we could define three kinds of laws:
    1) laws prohibiting hurting others (against murder, theft, robbery, etc)
    2) laws prohibiting ‘victimless’ behavior (against drug use, pornography, speeding-but-not-hitting-anyone, etc)
    3) laws which benefit the lawmakers at the expense of everyone else (against ‘subversive activities’, etc)

    I think types 2 and 3 correspond to the “bad laws”, and I think it’s plausible (though I haven’t data to back this up) that police attempting to enforce types 2 and 3 are much more likely to run up against restrictions intended to protect privacy than police attempting to enforce type 1 laws.

    If someone has been victimized, there’s someone to complain to the police, and point them in the general direction of the likely perpetrator. Even for crimes which by definition remove the victim’s ability to invoke the police (kidnapping, murder), there is usually a relative or friend of the victim who will take over that role. This lends itself well to procedures like search and arrest warrants which attempt to define when privacy can be relaxed for the sake of law-enforcement. Also, checking the population at random for murders/thieves/etc, in violation of privacy, is unlikely to be very effective.

    In contrast, if you want to enforce laws against drug possession, the police have to actively invade “private” space, since there’s no one complaining who will tell them where the drugs are. And searching the population at random for drugs, or monitoring passing cars for speed, is likely to be effective (in the sense of finding violators, not in the sense of eliminating drugs or speeding). Similarly for a police state looking for subversives.

    So I’d expect to find that the enforcement-damping effect of privacy is considerably greater on “bad” laws than on “good” ones.

    That said, I wouldn’t be surprised to find some bias at work in privacy’s popularity as well. I’m not sure the “bad laws” argument explains concerns about corporate collection of information at all, for instance.

  • Hopefully Anonymous

    The field of law in general is rife with cognitive bias, and is rather denuded of empirical basis for most law. So the recently accelerating emergence of empirical legal studies is a very good thing.

  • michael vassar

    It seems to me that the “bad laws” argument is one of the strongest arguments *against* privacy, as the elimination of privacy could be used to solve the coordination problems faced by people who violate bad laws. Currently, if everyone who, for instance, used illegal drugs, were to turn themselves in at once, the legal system would be swamped and presumably equal protection clause arguments could be used to invalidate those laws (since the alternative would be for the government to prosecute only those whom it had space to imprison). However, coordinating such protest is difficult. In a Transparent Society no-one would need to coordinate it. Someone could just turn in vast numbers of drug users themselves. A few people could solve the problem.

    It also seems to me that the potential benefits to social science strongly favor either a right to privacy or the elimination of such a right depending on weather social science is on net a public good or a public bad.

  • Douglas Knight

    michael vassar,
    That seems like incredibly naive faith in the legal system. By that standard, surely the equal protection clause is violated today. Perhaps total surveillance of the police could make it more enforceable.
    A similar argument I find more plausible is that currently untouchable drug users might lobby against drug laws. Surely they would feel more precarious under total surveillance and one major cost of lobbying–admission of drug use–would be gone.

  • Stuart Armstrong

    Alternative theory on privacy: it comes from the way we are brought up. A modern, middle class adolescent has his own room, and defines his independence through what he can do without parental approval or knowledge. We then carry these values into adulthood.

    Adolescent privacy is relatively recent; the obsession with privacy as well. A possible causation behind this correlation?

    There are societies, classes and epoques completely free of any privacy; this militates against us having an evolved bias towards privacy.

  • http://www.aleph.se/ Anders Sandberg

    If privacy was an evolved bias, we should expect to see it as a human universal. That does not seem to be the case. Looking at http://condor.depaul.edu/~mfiddler/hyphen/humunivers.htm
    I found just one clear privacy-related universal, that copulation is usually conducted in private (but historically certain copulations have been public, such as wedding nights in many cultures). That might be something evolved, but it doesn’t seem to be very strong. However, a real human universal is the tendency to compartmentalize our lives by location and social purpose, and it is not hard to see how certain compartments could get a privileged private status in some cultures. Especially if we have a weak “shyness bias” against appearing in vulnerable situations (this might not even be a bias, it is rational to reduce exposure when vulnerable). But cultural diversity in what is regarded as vulnerable is great; for example western ideas of the privacy and vulnerability of excretion seem to be relatively recent and local.

    I would expect privacy to be a concept that piggybacks on these weaker tendencies (compartmentalization and avoiding exposure when vulnerable). Imagine that we could not be hurt by others knowing what currently goes on in private. Would it still be a strong privacy presumption? It seems to lose a lot of power. However there would still be aesthetic reasons we might not want to know the details of what our loved ones do in the bathroom.

  • conchis

    In addition to Anders’ point, I don’t entirely see how the evolved privacy “bias” would constitute a bias. Rational behavior in prisoners’ dilemmas is not socially optimal, but it’s not biased either.

    I’m also attracted to hollywood freak’s point that, in part, our desire for privacy is a response to others’ biases (e.g. fundamental attribution error). Now, it may be that we’d collectively learn to avoid such biases more if there were less privacy, but it’s a difficult equilibrium to break.

  • http://profile.typekey.com/normansiebrasse/ Norman Siebrasse

    Anders and Stuart make an excellent point about the (lack of) universality in the privacy bias. If correct, this does not imply that privacy rights do not result from a cognitive bias, but rather that it ‘piggybacks’ on other tendencies, such as the salience bias, as I suggested in my post, or the ‘compartmentalization’ bias suggested by Anders.

    But I’m not giving up on a privacy bias per se. I’d like to have a specific example of a society completely free of any privacy. In particular, I have a hard time believing that there has every been a society in which couples having extra-marital affairs don’t try to conceal them, which is a different thing than simply not copulating in public.

  • http://profile.typekey.com/normansiebrasse/ Norman Siebrasse

    Bad laws / US Fourth Amendment: I did gloss over the ‘bad laws’ argument. In order to support privacy rights using this argument we have to show that privacy rights are more likely to effectively restrict bad laws than the alternative, which is the political process.

    The argument raised by thuddmonkey is that privacy rights primarily impede enforcement of bad laws. I quite sympathetic to this argument, which I call ‘privacy as procedural liberalism.’ If finding out what I am doing requires extraordinarily invasive measures, how can I be doing anyone any harm? While I find this to be extremely compelling in certain circumstances I think it fails as a general defence of privacy rights. Procedural liberalism doesn’t imply a general right of privacy, in the sense of a right to restrict flow of information about yourself, it implies a restriction on certain types of invasive searches. So, this view may justify the US 4th Amemdment, for example, which doesn’t give a general right of privacy, but is specifically directed to physical entry of the home (e.g. Wilson v. Layne, 526 U.S. 603, 610 (U.S. 1999); as TGGP put its, it is a right of property. (Ironically, in Bowers v. Hardwick, 478 U.S. 186 (1986) the US Supreme Court held that consensual homosexual activity between adults in one’s own bedroom, which seems to me to be the strongest possible example of an activity that should be protected by a right to privacy under the theory of procedural liberalism, was not constitutionally protected by a privacy right.) If we understand privacy in a broader sense of protecting information generally, rather than protecting certain physical spaces, it seems clear that many truly harmful activities benefit from privacy / restriction of information: tax evasion is private in that sense, as is shoplifting – any crime that requires ‘getting away’ with something. (The leading Canadian case on the right to privacy is a anti-trust investigation of a major corporation.)

    Further, one general response to the bad laws argument is that bad laws should be attacked directly, through the political process. Why should we think that privacy rights will be better at protecting against bad laws than the political process? Even though I am also very uncomfortable with criminalization of what I consider ‘victimless’ crimes, I have to admit that there are respectable arguments to be made for prohibiting drug use, pornography (and certainly speeding without hitting anyone), etc. I certainly can’t say the political process has failed just because I don’t always agree with the outcome. Nor is it clear that the courts will do better, as Bowers v Hardwick shows.

    A counter to the ‘use politics, not privacy’ argument is that the political process is broken, laws generally are bad, and privacy rights are a good thing even though – indeed because – they impede law enforcement indiscriminately. This is the ‘evil state’ argument. I don’t buy this argument because we have to compare the existing state with the realistic alternatives, not with impossible perfect worlds (what Demsetz called the ‘Nirvana fallacy’). I agree that there are many bad laws – -if you count them up, there might be more good than bad – but on the whole I’m confident that I’d rather have laws enforced, rather than the alternative of anarchy or dictatorship (unless it’s dictatorship by me). A variant is the argument that lack of privacy will lead to an evil state. I don’t buy this either. While I agree that evil states are unjustifiably intrusive, I think the causation runs the other way – evil states are a cause, not an effect, abuse of privacy.

  • TGGP

    Norman, it is possible that anarchy is a very bad option, but that on the margin we prefer less enforcement of laws given the current rather bad state. There are even some cases in which the government is bad enough anarchy is preferable! I refer to Somalia and the paper Better Off Stateless.

    If we treat privacy as a property right, doesn’t it matter hardly at all what variety of consensual activity is performed, with the important thing being that it is in one’s own home? Then banning contraceptives (as in Griswold vs Connecticut) is no more a violation than banning any other drug. Abortion is rarely performed in the privacy of one’s own home, but rather in facilities staffed by professionals, which would make it subject to the same power the government wields over other places of business and medical practices (whether that power is in fact legitimate is a whole ‘nother story!).

  • http://profile.typekey.com/normansiebrasse/ Norman Siebrasse

    I suppose it is theoretically possible that we have laws which are on average good but which are generally systematically over-enforced so that a general reduction in enforcement would be beneficial, but I can think of no particular reason to believe that this is true. Further, political recourse against over-enforcement is just as much an option as against substantively bad laws. For example, in the Canadian province of Ontario the introduction of photo radar dramatically increased enforcement of speeding laws. While many people complained on the obviously spurious basis that this was an invasion of privacy rights, photo radar was eventually eliminated purely as a result of political pressure.

    Certainly in some states anarchy would be better than the existing government. But in that case it is better to remove the government by revolution, rather than complain of invasions of privacy. And whatever should be done in Somalia, I don’t think this is relevant to the existence of the privacy presumption in countries such as the US.

  • agent00yak

    I would say that privacy is necessary for any large anonymous civil society due to signalling issues. In a small society where everyone knows everyone else, signalling isn’t as important, so privacy can be disregarded. However, in a largely anonymous society signalling is very important, as one bad impression can make someone decide they don’t ever want to deal with the person again. A reduction in privacy increases the time a person must spend properly signalling. Assuming that this signalling takes some effort and causes disutility, people in a large society would rationally prefer to consume some level of privacy.

  • http://profile.typekey.com/normansiebrasse/ Norman Siebrasse

    The comment by agent00yak, as well as the earlier comment along the same lines by Hollywood_Freaks, strike me as examples of the privacy presumption rather than explanations for it. Certainly we want to control the information that we communication to others. But by the same token, we don’t want others to be able to control the information that they communicate to us. So why is it that so many people presume that they will come out the worse if information flows more freely? Both commenters assume that others will systematically make misjudgments using the information they collect. Why? Do you believe that you systematically misuse your information about others? That you jump to conclusions and take information out of context? If you don’t, why do you assume that other people will?

  • TGGP

    However, in a largely anonymous society signalling is very important, as one bad impression can make someone decide they don’t ever want to deal with the person again.
    If it is very hard to avoid making a single bad impression, people might decide that is an overly strict standard. You seem to be indicating that the signal does not indicate much given its prevalence, so we would expect rational receivers of the signal to discount it to a certain degree.

  • Larry Knerr

    Do you believe that you systematically misuse your information about others? That you jump to conclusions and take information out of context? If you don’t, why do you assume that other people will?
    ‘Cause we read the news?

    The problem isn’t with “information flowing freely”, it’s with inequity of power. Given a situation where those with the power to do so could misuse your information for their own satisfaction or profit, why do you assume they wouldn’t?

  • http://profile.typekey.com/normansiebrasse/ Norman Siebrasse

    Larry,

    Certainly, others will use information about me to their advantage, if they can, just as I will use information about them to my advantage. Conversely, others will use their own private information to hurt me, just as I will use my private information to hurt them (cf. insider trading). This means that privacy rights hurt me as well as helping me. If I could get information about others while not revealing anything about myself, this would clearly be in my own interest. But privacy rights don’t just apply to me, they apply to everyone. That there is a generally held privacy presumption implies that most people believe that they will come ahead with stronger privacy rights / less flow of information. But this isn’t possible on average. This is the puzzle.

    If you re-read my comment, you will see that it specifically addressed to prior comments which suggested that most people would MISuse information, in the sense that they would draw WRONG inferences from information they receive. This is different from USING information against me, which means that other will draw CORRECT inferences and so gain an advantage. If it were true that most people will draw WRONG inferences, this would indeed be a good reason for the privacy presumption. People would be harmed by releasing information about themselves (which is clearly true whether or not others MISuse that information), but if I MISuse information about others, then I do not gain any offsetting advantage. While I certainly do believe that others will USE information about me to their advantage, I see no reason to think that I, or they, will MISuse that information by systematically drawing INcorrect inference.

    With respect to your point about inequity of power, the term “power” needs to be more clearly defined: simple appeals “power imbalance” and related terms are notoriously conclusory. One imbalance of power is in the state vis-avis the citizen, where the state has powers of search etc. that the citizen does not have. This is why the ‘bad laws’ argument is quite compelling. This is discussed in more detail above. If by “power” you mean people with more information, the point is circular. It is true that people with more information will exploit those with less information, but fewer privacy rights would minimize the information imbalance. If by “power” you mean “wealth” I do not agree that poorer people systematically derive more benefit from privacy laws than richer people. On the contrary, my strong intuition is that poorer people generally have less to hide, and richer people generally have more access to private information and more opportunity to use it than do poor people. Indeed, if I were a conspiracy theorist, I might suspect that the media attention to privacy rights is part of a plot by wealthy media barons to protect the secrecy which helps them maintain their power.

  • Hopefully Anonymous

    Norman,
    Very clever post (particularly the ending). I appreciate that you’re bringing nuance into this whole powerful/powerless strain of discussion. As if the world (or American) population could be so easily sorted into those two categories.

  • Larry Knerr

    I’m sorry for misreading you, and hope I’m not doing it again.

    That there is a generally held privacy presumption implies that most people believe that they will come [out] ahead with stronger privacy rights / less flow of information. But this isn’t possible on average. This is the puzzle.
    Doesn’t this confuse global rationality and local rationality? What makes sense for me doesn’t have to make sense for all of us collectively. Insofar as private data turns into a commons when made public or political, then we really will come out ahead by not playing. (Similarly with law, as others mentioned.)
    And on rereading, I see you pretty much said just that. Though I still don’t think that’s a cognitive bias in favor of privacy so much as a sensible (sensibly biased?) mistrust of government.

    Isn’t it reasonable to suppose that where people desire and can afford to be ignorant and biased – that is, where reality doesn’t provide feedback – they often choose such beliefs? Much of the data-mining and related privacy objections concern new information which has heretofore not been subject to cultural and legal restraints, and everybody from the “information wants to be free” folks to the government seem to want to go the “all your data are belong to us” route for others’ used-to-be-private info, and the path of wailing and gnashing of teeth and complaining to their representatives for their own.

    If the state had equitable laws, evenly applied, in information at least as much as elsewhere, then people could trust it to protect them from abuse of privacy. As Solove (TGGP’s link) points out, this is hardly the case. (Or we could ask a “wealthy media baron” like Conrad Black how much power and secrecy he has, hey?) Instead, we have something that looks to me like an un- or mis-managed commons, and instead of trying to find a way to manage it by privatization or regulation, government mostly acts as one of the more abusive of the players trying to profit – either directly, e.g., for bureaucratic convenience, or indirectly, e.g., laws traded for campaign funds. Is it biased to be mistrustful?

    One last possibly-relevant bias for your collection – I don’t know a name, but call it “control” bias – the illusion that you’re safer driving a car than riding in a plane because the decisions are in your hands. Again, not necessarily irrational when attempting to protect personal information from abuse in the absence of sufficiently good legal or cultural protections, but biases are rational in some situations or we wouldn’t have evolved most of them.

  • Hopefully Anonymous

    Larry Knerr, very good naming and identification of a bias (control bias), and I think a major one. I haven’t read about it before.

  • Norman Siebrasse

    Larry,

    Thanks for persisting. There are some problems with the bad laws / bad state argument in favour of privacy (see my July 19 comment), but enough people have raised it that I think this must be the heart of the problem. I am going to have to think about it more carefully before I can reply fully.

    Incidentally, given your emphasis on inequity of power, I think it’s interesting that you’ve chosen the Conrad Black example to illustrate why we should distrust the state to apply laws equitably. I would have thought it is a good illustration of my point that privacy rights are likely to benefit rich people disproportionately. Most people (well, me anyway) don’t have enough money to make nefarious dealings worthwhile, so they have nothing to fear from full financial disclosure (apart from mild embarrassment over how little money they have), while people like Black are much more likely to be in a position to use their wealth and connections to illegal benefit. Ordinary people are likely to be in the position of shareholders in the company plundered by Black. So far as I can tell, the law was applied equitably, at least in that Black was convicted despite being rich. (I haven’t followed the case closely enough to have a substantive view on the merits of the conviction.)