Follow

Privacy is not a science, it is a human right

ar.al/2019/02/14/privacy-is-no

My response to a dangerous turd of an article by Bart van der Sloot (co-founder of Amsterdam Privacy Week – sponsored by Palantir and Google – and docent teaching “Privacy and Big Data” at Tilburg University)

Mr. Sloot attempts to reframe privacy as a “science” and shame those who work to protect it for having a “pro-privacy agenda” & not being “neutral”. He also calls for “new rules” to disallow such biased behaviour.

@aral ooohhh myyy goddd what a scumbag. what a completely soulless individual. to try to argue that privacy isn't scientifically valuable. what a straight up evil thing to say. this dude gets the wall. wow thats infuriating

@aral
I have heard this kind of thing before: privacy is a recent idea; privacy is not a right; privacy is outdated as a concept. All from social media "scientists" (coincidentally working for google or fb 🙄)
None of it stands up to a moment's scrutiny.
The fuckers.

@BadCarrot
It's a creepy, disingenuous and sick way of decentering human rights.

It's evil at work. It needs to be called out for what it is, and resisted.
@aral

@aral Wow what a shithead! Course I'm not surprised, given how Zuck says "Privacy Is Dead". We really need more privacy advocates like you out there being heard by the masses. Ones that don't take money from the Capitalists. It's a hard road but someone has to do it. There are way too many slimeballs in positions of power after all.

@aral Hm, I don't fully agree with your position that privacy is not a science. Privacy is a human right, of course, but that does not mean that it cannot also be treated scientifically.

Psychologists can research the need for privacy, sociologists can research its social implications, computer scientists can study technical means to ensure privacy.

I do agree, however, that scientists researching on privacy don't need to be neutral on the topic.

@colomar How do you « treat scientifically » privacy? How do you reliably reproduce experiments ruling out external factors about privacy, which is the only particularity of science?

@hypolite
I think privacy can be approached scientifically--anything can--but it does not need a scientific basis to be valid.

Any approach to it should center human beings, their demonstrated needs, and their dignity, or it's less a science and more a tech-based violation of people's personal integrity.
@colomar

@erosdiscordia Same question, how do you approach anything scientifically without needing a scientific basis? What are you even validating and how?

@hypolite
I think I agree with you. I'm just saying that science can gather data around a topic like privacy (or parenthood, or love, or religion, or creativity), but the usefulness of that data will be contingent on whether the ineffable human component of the concept is centered. The science can only complement the reality of the issue. Not define or control it. To try otherwise is a type of violence.

@erosdiscordia Ok, sounds like the classic confusion between "scientifically" and "statistically". Privacy (or anything) can indeed be studied statistically, but not really scientifically.

Like you said, privacy advocates can and will take sides, either in favor or against privacy. There's no neutral point from which privacy can be safely studied.

@hypolite
Ok, I stand corrected on the terminology then! And yeah I definitely agree, there's not really a neutral position to study it from.

@hypolite The same way that you do other psychological or sociological research. Why would privacy be fundamentally different from, say, safety or security?

@colomar Because I would argue that "treat scientifically" is a misnomer in all these cases. "treat statistically" probably would be better suited, as they aren't the same although often confused with each other.

@hypolite
"Science (from the Latin word scientia, meaning "knowledge") is a systematic enterprise that builds and organizes knowledge in the form of testable explanations and predictions about the universe."

I would say that testable explanations and predictions about privacy are indeed possible. One can e.g. conduct proper scientific experiments on what people perceive as an invasion of privacy or not.
Unless you think psychological experiments are not possible.

@colomar Not in the scientific meaning I stated above (reproducibility and external factor reduction). Because each human is different, you won't get precise consistent results across ages, genders, cultures, geographical locations, etc...

It doesn't mean that psychological experiments aren't possible at all, just that their results and their interpretation are open to way more scrutiny because of this lack of precision.

For example, German people are more sensitive to privacy than any other country I know about. It is a statistical (or empirical) fact with no scientific basis. However it probably has historical and cultural basis.

@hypolite Have you read through some literature covering psychological experiments? Do you know how they're done?
Significant care is taken to either eliminate external influence or use a big-enough, representative sample to randomize it out when comparing multiple experimental conditions.

Objectity, reliability and validity are the three main criteria for the quality of psychological experiments, just like in other empirical disciplines.

@colomar Yes, I know about the statistical tools, the sample size constraints, the error margin, but that's my point, they're all statistical tools, not scientific tools where external factors are removed a priori, no manually culled from the data.

I believe there is a fundamental difference between having the most relevant sample size and reducing an experiment to the only relevant factor. Because in the latter, the margin of error is multiple *orders of magnitude* smaller.

@hypolite
So biology is not science then, either? They use statistics to cope with randomness in their experiments as well.

@colomar The issue isn’t “using statistics” in general, it’s “only using statistics to get results”.

In biology you can physically isolate or reduce a subject to reduce the margin of error before using statistics, but for psychology it isn’t possible (nor ethical).

@hypolite
Of course it is both possible and ethical in psychology, depending on what you want to research.
Most psychological experiments happen in a lab, under strictly defined conditions, in order to eliminate or control as many external influences as possible.

It is not "all statistics" like you seem to think. Statistics are only used to deal with any remaining sources of influence that cannot be controlled. Controlling conditions is always preferred.

@colomar I’m sure! But you can’t isolate a feeling for example, nor an individual far from any exterior influences for too long.

Ad I’m not dismissive of psychology as a whole, nor social science for that matter (even if I still think it’s a misnomer), but the error margins simply are in a different league than those used in “hard science”.

And I believe this common confusion is exploited by van der Sloot who infers science’s infinitesimal error margins can be somehow applied to privacy studies, which would indeed render their results “neutral”, but we’re very far from it.

@hypolite @aral It is important to distinguish between basic and applied research: When I was a an academic researcher in usable security and privacy, I would have been fine with doing a study on how to design privacy controls in a social network in an understandable way sponsored by FB or Google, because in that case their interest is the same as the users'.
For a study about the psychological implications of surveillance, however, I'd stay clear of both.

@colomar Thanks for the insight about basic (theoretical ?) and applied research. I think I was doing the separation anyway: climate change is a scientific ("neutral") fact that comes from basic research, while what we should do about it is applied research.

@hypolite
Btw, for context: I have a German Diploma (equivalent to MSc, and yes psychologists get an MSc not an MA) in psychology plus three years as an academic researcher, so I have conducted lots of psychological experiments.

@colomar For what it’s worth, I never doubted your legitimacy or knowledge.

@hypolite
What you seem to be thinking of are correlative studies, which are indeed only based on statistical correlations between naturally occurring phenomena.
Those are the least preferred method in psychology, however.
Usually a psychological theory is only widely accepted if its hypotheses can be confirmed using experiments with manipulated conditions.

@colomar Yes, they are the one who get a lot of press with misleading titles like “X causes Y, study says”.

@colomar @aral
"scientists researching on privacy don't need to be neutral on the topic."

They should try to be neutral while gathering and processing data, i.e. not compromise their methods, but they also should see it as a moral imperative to use their insights to further human progress, i.e. preserve & restore privacy.

Science is never non-normative, & all scientific endeavors need to sit on a defensible ethical framework.
(Logically, all science follows from philosophy, cf. epistemology.)

@aral Reminds me of that one scientist sponsored by Big Oil who publicly denied and ended up creating a public "debate" on climate change.

The same old tricks.

@aral Aral, believe it or not, I think you might actually be _under_selling just how bad this is!

Nobody in their right mind would ever say “Having conferences about ethical genetic science suggests that there are biases with respect to selectively breeding and cloning humans, so that’s bad science.”

@aral The reason why they want to describe it as a science is so that they can then pedantically argue over details where you'll get bogged down with precisely trying to define things. "That's not the correct definition of privacy", they'll say, "which peer-reviewed journal supports your view? Where is the formal proof of your assertion?" and things like this.

They'll probably hire some mathematicians to get some "privacy science" formal proofs into a vanity journal, preferably alongside more credible cryptographers.

I can guess because I've come across this type of personality before in other contexts. It's possible to dress up prejudices in math and then pretend that they're "neutral", "formally validated" or "backed by the scientific community".

"Are you trying to refute all of mathematics?", they'll say indignantly when you complain that their article is BS.

@bob Nailed it.

Attempting to academically disqualify almost everybody from a conversation about a topic that affects literally everyone is a misuse of the idea of science.

It's getting between people and themselves. It's a control-grab.
@aral

@aral Science is probably invading people's privacy while being legally OK.

@aral
This came as a surprise. I don't know van der Sloot's work that well but enjoyed some parts of his article on privacy interpreted as non-domination in courts.

Seems like they are building a similar divide that we have in futures studies (emancipatory research / instrumentalist foresight) and also in security research (peace research / traditional military security research). Can they really not see how this dismissal of ethics is dangerous with privacy at this point in time?

Sign in to participate in the conversation
Aral’s Mastodon

This is my personal Mastodon.