Circling the Drain in the NSA Surveillance Debate

Avatar

Ethan Gach

I write about comics, video games and American politics. I fear death above all things. Just below that is waking up in the morning to go to work. You can follow me on Twitter at @ethangach or at my blog, gamingvulture.tumblr.com. And though my opinions aren’t for hire, my virtue is.

Related Post Roulette

55 Responses

  1. Avatar Mad Rocket Scientist says:

    Thank you Ethan. Excellent post!Report

  2. Avatar Chris says:

    What’s more remarkable than Simon’s faith in the undemonstrated success of these counterterrorism policies is that he thinks the abuses will be any more egalitarian than the abuses committed by any institution.

    I believe this was Jason’s point, and Burt’s, in their initial posts on this subject here.

    On the one hand, I understand Simon’s point about the wake up call. While I’ve known a lot of people on the left who’ve been upset about racial profiling, a level of scrutiny that goes way beyond collecting metadata, the level of outrage we’re now seeing in some circles because of revelations that data is being collected on everyone, including them, far outpaces any outrage they’ve shown in response to racial profiling. So it does seem a bit… if not hypocritical, then selective. “Oh, you mean they’re collecting data on me? Well, now I’m outraged!”

    On the other hand, it seems to me that if racial profiling is wrong for reasons beyond simply that its being discriminatory — that is, if it’s also wrong because treating innocent people like criminals is wrong — then expanding it, even in a limited way (and I hope we can agree that collecting metadata is far less intrusive than, say, repeatedly stopping and frisking people for no reason other than that their skin is too dark or they’re in the wrong neighborhood) is also wrong. What’s more, the expansion of its coverage doesn’t imply, as you, Jason, and Burt note, the expansion of its abuse.Report

  3. Avatar Barry says:

    “What’s more remarkable than Simon’s faith in the undemonstrated success of these counterterrorism policies is that he thinks the abuses will be any more egalitarian than the abuses committed by any institution.”

    Hmmmmmmmmm…………..faith in undemonstrated success………why do I think ‘War on Drugs’?Report

  4. Avatar trizzlor says:

    What’s more remarkable than Simon’s faith in the undemonstrated success of these counterterrorism policies is that he thinks the abuses will be any more egalitarian than the abuses committed by other institutions.

    Surely it will be the poor, racially and ethnically marginalized that suffer the most under an ever expanding surveillance regime, just as they suffer more under nearly every other policy.

    Simon’s point is that the big-data approach to law-enforcement is just as likely to lead to more egalitarian outcomes, and your vague reference to the disparate outcome nearly every other policy doesn’t really rebut this point. As Simon originally argued, the risks for abuse that you reference are already present in any law enforcement policy. But where PRISM-type programs differ is in replacing much of the biased human element with hypothesis-free data mining. In the scenario that Simon describes pre-PRISM, a detective would have to dig through the scraps of publicly available information on the suspect’s social circle to prioritize potential leads. How high is the likelihood that someone from a marginalized group – a recent immigrant, an Arabic sounding last name, attendance at a Mosque – will be unfairly targeted? I’m sure detectives working under pressure with limited evidence take this kind of profile-based shortcut all the time. With PRISM, on the other hand, all available social network information is analyzed for trends and compared against a massive database of benign individuals. Dumb profiling based on name or appearance won’t even factor in, and the content of the conversations will be weighed much more heavily than relatively benign indicators like ethnicity or religion or income. It is not obvious to me that the latter system introduces more bias, and quite likely that it does the opposite.Report

    • Avatar Mad Rocket Scientist in reply to trizzlor says:

      I think the crux of programs like PRISM falls to two questions:

      A) Who has access to the database
      B) How is the data accessed?

      Limit access to LEO, and bar political access, and at least you can limit the potential abuses to a degree.

      Construct a database that can only provide data when an algorithm pings to a threat, rather than one that can be scanned by name, ethnicity, etc – that way the database can not be used by someone with an axe to grind against an individual.Report

      • Avatar trizzlor in reply to Mad Rocket Scientist says:

        Exactly. Anonymizing this system would be a peace of cake. Heck, make it so the thing can only ping once you’ve gotten a warrant for the individual, forcing it to be a secondary information system. The problem is that even after Mr. Gach argues that each tool presents unique benefits and costs, he fails to consider the serious social benefits of such an automated system can provide, particularly to traditionally marginalized groups.

        Not to get too deep into the psychoanalysis, but I think part of the problem is that if you asked various people the following question: “You’re in an airport screening line and are getting profiled for a “special” screening, would you rather the profiler be a random TSA agent who gets to see your passport or a soulless computer that gets to see your passport and your cellphone metadata?” you’d get a very different answer depending on who you ask. I’m quite sure my Sikh co-workers – who “randomly” get screened every time they fly – would pick the soulless computer, but they’re not the ones driving this conversation.Report

      • Avatar Barry in reply to Mad Rocket Scientist says:

        “I think the crux of programs like PRISM falls to two questions:

        A) Who has access to the database”

        I’m sorry, but that’s classified.

        “B) How is the data accessed?”

        I’m sorry, but that’s classified.

        “Limit access to LEO, and bar political access, and at least you can limit the potential abuses to a degree.”

        Access is limited to Cleared, Pure Americans with Pure Motives. Anybody who tells you otherwise is lying, and we have a 99% redacted document to prove it.Report

      • Avatar zic in reply to Mad Rocket Scientist says:

        I think there’s a third, and that’s what data is being collected.

        Does it, for instance, contain cell-phone pings to a cell tower, and chart travel via switching towers? Meta data can mean a lot of things. We were mislead about our bank records.

        We were mislead about the privacy of overseas phone calls. Now, we’ve been mislead about the sum total of our communications via cellphone and internet. Because there is a very disturbing pattern of collecting as much as possible and 1) keeping the actions secret, and 2) lying about the actions.

        This now spans two administrations, so I have to presume much of what we’ve witnessed is the culture of the natural security apparati.Report

    • Avatar Barry in reply to trizzlor says:

      “But where PRISM-type programs differ is in replacing much of the biased human element with hypothesis-free data mining. ”

      First, any action taken by people will be filtered through people. It’s likely that various human parties will be tasked with sifting through the automated output, and deciding which things to follow up on, and how.

      Second, people will program those algorithms. And adjust them.

      Third, those algorithms will not necessarily start ‘unbiased’.

      Fourth, the costs of false positives will be least when the victims have lower power, status and money. That will feed back into (1) and (3).

      Fifth, this all assumes that the program will not be quite deliberately steered in the sort of directions that many criminal ‘justice’, ‘national security’ and police programs have ended up going.Report

  5. Avatar MtVernonCannabisFarms says:

    We need to have this debate … right , after 7 years of the decision secretly dictated .
    Everybody’s changing the subject , secret laws , secret courts , secret decisions .
    Means and methods of doing things you can’t talk about because they can’t release the details .
    Not even possible to debate the limits of what can’t be mentioned . Its a joke thats going to boil down to blind obedience or blind scepticism

    one hell of a way to run a countryReport

    • Avatar Damon in reply to MtVernonCannabisFarms says:

      That’s why I’m for ending it. Full stop.

      Any power someone has will eventually be abused. We see this all over the gov’t. The only way to prevent abuses is to eleminate the acess to the data.Report

      • Avatar Ethan Gach in reply to Damon says:

        This is the key point. All government programs will be abused–the question is whether the benefits outweigh the abuses.

        I’m of the mind, for instance, that whatever abuses to social security disability, or TANF, the benefits are too great to get rid of the programs.

        With counterterrorism surveillance though, I’m extremely skeptical (i.e. have no reason to believe) that that is the case. Until the NSA is able to show me proof of why I need them on that wall, I’m for getting them off of it before they, to use Simon’s analogy, accidentally shoot someone.Report

        • Avatar Barry in reply to Ethan Gach says:

          “This is the key point. All government programs will be abused–the question is whether the benefits outweigh the abuses.”

          And what controls, follow-up, auditing, etc. are available. With these programs, the high-end guess would be ~1% of the controls, follow-up, auditing, etc. that other government programs get.Report

  6. Solid post, Ethan. You’re killing it on this topic.Report

  7. Avatar Stillwater says:

    Ethan, Here’s a link to an important article on the origins and development of the PRISM including the history of the NSA’s meta-data collection methods, and some of the legal questions surrounding it. I think it’s an important contribution to the overall debate, but of course, you be the judge.

    NSA uses ‘terrorism’ to justify mass surveillance that started long before 9/11 and the Patriot Act.Report

  8. Avatar CK MacLeod says:

    Sorry, y’all, that I couldn’t participate in and deconstructively divert the above colloquy on Searle and Derrida. Am thinking about how to continue the discussion if and when I can. Worse comes to better, I can always host it at my own blog, I dunno, sometime in the next days or weeks, unless of course the LOOG Powers that Be, Whoever They Are, don’t do the right thing in the meantime and give us a sub-blog dedicated to this purpose. As for doing it up as a post or comment rescue or whatever, I’ve never been initiated into the secret handshakes and so on, so don’t know anything about that. IMO it would be worth doing, and as opaquely as possible, even if only to dismay Hanley.Report

    • Avatar Stillwater in reply to CK MacLeod says:

      Either way, CK. It’d probably be better here, since this is where we already are.

      If not, I’ll come over to your place.Report

      • Avatar CK MacLeod in reply to Stillwater says:

        Well, as I said, Stillwater, I am not currently in a position to make anything happen here except by hijacking a thread somewhere, possibly in the shadow of a merciless 4-day thread-shutdown rule. As is typical for this “place,” unless you’re in the club, it’s a complete mystery how all such decisions on FP access, sub-blog privileges, etc., are made, or even whom to contact or how to do it. In fact, it’s such a deplorable example of non-transparency and non-accountability that I’m not sure I could justify giving it an implicit endorsement by authoring a post, which in the LOOG’s implicit class system is very different from being a mere commenter.Report

        • CK,

          The reasoning for the shutdown rule is that, well, this place is administered by volunteers and it becomes a drain on our time and resources when threads spiral out of control. Threads that last more than four days do so at a rapid clip (the longer a thread goes, the more likely it is that it has devolved into insults or the repeating of previous points, but more acrimoniously and at a higher volume). A uniform standard, consistently applied, comes across more fair than 100,000 questions about why we shut down this thread but not that thread, and why we shut it down precisely when so-and-so got the last word.

          I can understand your frustration. While perhaps an announcement would have been ideal, it changed a number of times until the right number of days to leave threads open was found.

          I would not worry about baggage that comes with being a guest post author. We maintain a reasonably liberal policy in that regard. It isn’t an indication that we favor or disfavor someone. A couple months before being invited to be a contributor, I had a guest post proposal that was denied. On the other hand, relatively new commenters and people who rarely comment at all have been given the floor.Report

Leave a Reply

Your email address will not be published. Required fields are marked *