Extended Thoughts on HumanID and BI/OS
Jul. 1st, 2009 05:41 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
wolven7Please remember our earlier discussion of the BI/OS and the Human Identification at a Distance programs.
Please couple this with whatever it is that you think you understand about myself, my family, and my interest in cybernetics and biotechnologies, as well as my Something Amazing About Myself, yesterday.
Let's be perfectly clear: I don't 'scrutinize gait," when I observe and identify people at a distance. What I do is observe and discern the probability of a certain type behaviour, based on past experience with the person identified, and human beings, as a whole. (David Bowie - [Leon Takes Us Outside]). I then add that probability to the list of potential things that may be about to happen to or be done by them, to or by me, to or by others in my environment, &c. (David Bowie - [Outside]). What this means is that I look at a situation, and I factor in everything i know, feel, and intuitively believe (and that's arational intuition, no matter what neurophilosophers try to tell themselves) about every thing in the field of consideration. It is a robust blend of emotive symbolism, pattern recognition, statisitcal modelling, and inductive reasoning. In short, it's what everyone does, every day, to get from point A to point B.
Now, with all of that said, it may be surmised that I would think that this should be expanded and used by DARPA's HumanID/BIOS programs to track and predict potential threatening bahviour. (David Bowie - [The Heart's Filthy Lesson])But this surmisement (surmisation? Surmise) would be wrong. I do not necessarily believe that these programs would be good things, either individually or combined, and here's why:
There is a correlation algorithm that pairs certain types of stance, posture, moviement, voltaic skin and involvutary respiratory and cardiac fluctuations with certian states of mind, and trains of thought. (David Bowie - [A Small Plot of Land]). In controlled experiments, these have been shown to be fairly decent indicators for the factors sought. The problem, however (and I think it's a major one), is that their probability algorithms aren't flexible enough, yet, to adapt to the possibility of fleeting thoughts, trouble at home, &c. In other words, they have not been expanded, enough to take into account the effects of various brief, arational, physio-chemical, emotional combinations that may all add up to a certain response in one person, but not another. On top of this, we have the danger of confirmation bias, where we may have sought specific answers for specific questions.
So: Independently, the algorithm isn't "human" enough to predict arational, symbolically-influenced behaviour patterns. In conjunction with human operators, an over-arching envorionment of fear as well as a potentially unconscious need to punish or scapegoat someone matching a specific symbolic, arational resonance could easily lead to wrongful persecutions. (David Bowie - [Segue- Baby Grace (A Horrid Cassette)]). And we won't even get into the eventual backlash against a perceived wave of such persecutions. So, until humanistic psychological Possibilities can be better correlated with the algorithmic Probabilities, we need to limit the likelihood that we will stop for extra screening, strip search, and destroy the personal belongings Of a person, because they, essentially, "Looked Suspicious."
David Bowie - [Hallo Spaceboy]--- We must remember that this-- the effort to accurately model someone's behaviour, Prior to Their Decision Making-- is territory that we still deeply misunderstand. fMRI studies point in particular directions, regarding decision making and agency, depending on the Prior Information Given To The Subject and Operators, Alike. This means that, should we inform the individual that they are going to be under observation, of a certain type, they will be A) more careful about their behaviour, B) Too careful about their behaviour, C) Generally Outraged And Nervous, and/or D) Any combination of the above, plus others I can't think of, right now.
We don't know enough about what our symbol sets mean to ourselves, and their effects on our physical states, let alone what those of other people do to them, to try and accurately predict that which is, in essence, a merely potential act.
David Bowie - [The Motel]--- You can call it, "Precrime," "thoughtcrime," "sense-offense," or whatever else you want. What it boils down to is the distinct and onerous possibility of the misapplication of tools that we misunderstand, or underestimate, unless we recognise the limits of our understanding, try to correct and allow the space of flexibility for them.
David Bowie - [I Have Not Been to Oxford Town]--- Thoughts?
Please couple this with whatever it is that you think you understand about myself, my family, and my interest in cybernetics and biotechnologies, as well as my Something Amazing About Myself, yesterday.
Let's be perfectly clear: I don't 'scrutinize gait," when I observe and identify people at a distance. What I do is observe and discern the probability of a certain type behaviour, based on past experience with the person identified, and human beings, as a whole. (David Bowie - [Leon Takes Us Outside]). I then add that probability to the list of potential things that may be about to happen to or be done by them, to or by me, to or by others in my environment, &c. (David Bowie - [Outside]). What this means is that I look at a situation, and I factor in everything i know, feel, and intuitively believe (and that's arational intuition, no matter what neurophilosophers try to tell themselves) about every thing in the field of consideration. It is a robust blend of emotive symbolism, pattern recognition, statisitcal modelling, and inductive reasoning. In short, it's what everyone does, every day, to get from point A to point B.
Now, with all of that said, it may be surmised that I would think that this should be expanded and used by DARPA's HumanID/BIOS programs to track and predict potential threatening bahviour. (David Bowie - [The Heart's Filthy Lesson])But this surmisement (surmisation? Surmise) would be wrong. I do not necessarily believe that these programs would be good things, either individually or combined, and here's why:
There is a correlation algorithm that pairs certain types of stance, posture, moviement, voltaic skin and involvutary respiratory and cardiac fluctuations with certian states of mind, and trains of thought. (David Bowie - [A Small Plot of Land]). In controlled experiments, these have been shown to be fairly decent indicators for the factors sought. The problem, however (and I think it's a major one), is that their probability algorithms aren't flexible enough, yet, to adapt to the possibility of fleeting thoughts, trouble at home, &c. In other words, they have not been expanded, enough to take into account the effects of various brief, arational, physio-chemical, emotional combinations that may all add up to a certain response in one person, but not another. On top of this, we have the danger of confirmation bias, where we may have sought specific answers for specific questions.
So: Independently, the algorithm isn't "human" enough to predict arational, symbolically-influenced behaviour patterns. In conjunction with human operators, an over-arching envorionment of fear as well as a potentially unconscious need to punish or scapegoat someone matching a specific symbolic, arational resonance could easily lead to wrongful persecutions. (David Bowie - [Segue- Baby Grace (A Horrid Cassette)]). And we won't even get into the eventual backlash against a perceived wave of such persecutions. So, until humanistic psychological Possibilities can be better correlated with the algorithmic Probabilities, we need to limit the likelihood that we will stop for extra screening, strip search, and destroy the personal belongings Of a person, because they, essentially, "Looked Suspicious."
David Bowie - [Hallo Spaceboy]--- We must remember that this-- the effort to accurately model someone's behaviour, Prior to Their Decision Making-- is territory that we still deeply misunderstand. fMRI studies point in particular directions, regarding decision making and agency, depending on the Prior Information Given To The Subject and Operators, Alike. This means that, should we inform the individual that they are going to be under observation, of a certain type, they will be A) more careful about their behaviour, B) Too careful about their behaviour, C) Generally Outraged And Nervous, and/or D) Any combination of the above, plus others I can't think of, right now.
We don't know enough about what our symbol sets mean to ourselves, and their effects on our physical states, let alone what those of other people do to them, to try and accurately predict that which is, in essence, a merely potential act.
David Bowie - [The Motel]--- You can call it, "Precrime," "thoughtcrime," "sense-offense," or whatever else you want. What it boils down to is the distinct and onerous possibility of the misapplication of tools that we misunderstand, or underestimate, unless we recognise the limits of our understanding, try to correct and allow the space of flexibility for them.
David Bowie - [I Have Not Been to Oxford Town]--- Thoughts?
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
no subject
Date: 2009-07-02 02:16 am (UTC)"Won't someone please think of the children???"
I'm much more for having very severe punishment for crimes vs. endless slaps on the wrist for doing nothing at all except fitting a profile, which seems to be where we're going...
Unless you download an mp3... Then it's consecutive life sentences while some rapist is set loose in your neighborhood.
no subject
Date: 2009-07-02 04:05 pm (UTC)I think I wrote something similar last year...
no subject
Date: 2009-07-02 08:25 pm (UTC)no subject
Date: 2009-07-02 08:23 pm (UTC)no subject
Date: 2009-07-07 02:11 am (UTC)no subject
Date: 2009-07-07 03:13 am (UTC)All we need is a rogue AI, and a thing I saw in a prophetic trance, 8 years ago, is that much closer to coming true.