By Daniel Rigney
“Calculations are in the driver’s seat, and they drive us.”
Many ages ago, as a graduate student in sociology, I enjoyed referring to quantitative social research as the “dance of the variables.” We apprentice researchers were then learning to work with what are now called 'metrics,' or measurements of variables, relating to class, ethnicity, gender, education, religion, and the like. We were learning to feed these data through various statistical procedures to produce precise-looking correlation coefficients, regression lines, and significance levels. The resulting tables and graphs looked impressively objective when inscribed by a machine, in dot-matrix characters, onto perforated computer paper.
At the time, it amused me to imagine a researcher racing back from the computer center with the exciting news that “mothers’ religiosity explains an additional 3% of variance in son’s status attainment!” Be still, my heart.
I shouldn’t have been so glib. Quantitative social researchers produce data and analyses that are essential to the survival and sustainability of modern information societies. Their role in the today's world is indispensable. Yet there are dangers lurking in the practice of transmuting lived human experience into remote mathematical abstractions. It is as though society itself were now a Pythagorean construct – a matrix of numbers made flesh, or rather flesh made numbers. It's as if the metrics and models, quantitative representations of social reality, are more real and significant than the lives of the people from whom they derive.
These days, the dance of the variables is evolving in an ominous direction as ‘quants,’ not just in finance but across many other fields, devise increasingly abstract and derivative metrics, models and 'instruments' that drive corporate and governmental policy, derivatives that can be crushingly real in their human consequences.
Call it the dance of the algorithms. Algorithms are those step-by-step calculating procedures that are now making many of our most consequential decisions for us, automatically and robotically, in fields ranging from finance to education, and from political campaigning to sports. (Did anyone happen to see Moneyball?)
Algorithms are like mathematical machines, performing stepwise computations that determine the fates of real people's lives. Here are just a few of the algorithms that now seem to be steering our destinies in the early 21st century:
Wall street financial instruments, including financial derivatives, that can behave in unexpected ways as they collide with other variables in the real economy (not to be confused with economists' models of the real economy, which, weirdly enough, have themselves become variables in the real economy);
Googlish search algorithms, the secret proprietary formulas that determine which search engine optimization (SEO) strategies will fail or succeed, and thus which memes or ideas will die or live on to reproduce;
School funding formulas, and especially those based derivatively on high-stakes testing, that determine which kids will receive adequate educational resources and which won’t;
Political algorithms used by campaign strategists to separate the signal from the noise in economic and demographic data, and thus to calculate which messages to send to whom, when, and by what channels in order to maximize fundraising and voter participation on behalf of this or that political commodity. (See also market research.)
Exam problem: Add two fresh examples of your own.
Not only do programmed algorithms make many of our decisions for us, but they also interact (‘dance’) with each other in unexpected ways that far surpass our capacity to predict or restrain their interactive effects.
The derivatives that Wall Street mathematicians concocted prior to capitalism’s last near-fatal crash in 2007-8 played out in ways that even their makers failed to anticipate. Other algorithms across society are now propelling us in uncharted directions that no one fully fathoms. Like all social inventions, they can have unintended consequences for good or ill.
Algorithms have now joined the dance of the variables, becoming social factors unto themselves. Mathematical procedures are coming to replace conscious human decision-making with automated decision-making in a widening array of social institutions.
I’m not suggesting that all algorithms are malignant. On the contrary, some may do immense good. I, for one, would be happy to see a politically neutral algorithm (if such were possible) designed to draw fairer legislative redistricting lines, to replace the current DeLay System, which operates intentionally and effectively to disenfranchise millions of Americans.
Algorithms can be helpful tools. Yet they may also inadvertently create mathematical disasters equaling or exceeding the destructive force of natural disasters such as hurricanes. We’ve already seen the effects of one such disaster in the past decade, the role of quants and derivatives in the financial cataclysm of 2007-8, responsible in some part for the near collapse of the global economy.
Now comes the really scary part. Our multiple algorithms may interact with each other, directly or indirectly, to produce inadvertant Superalgorithms (or more accurately, networks of discrete algorithms) whose potential combined effects are profoundly beyond our capacity to predict or control, like mathematical frankensteins, or viruses escaping the lab and hybridizing in the wild. Imagine, for instance, how stock market algorithms might interact with taxation algorithms, affecting revenues and expenditures in algorithm-driven social service and educational programs.
Even the smartest people in the room (academic, governmental, military or corporate) may be unable to save us from the consequences of these accidental interactions among mathematical machines. Indeed, the smartest people in the room may well be the very ones who unleash them into the world. And like the sorcerer's apprentice, they may find that the forces they unleash have an inhuman life of their own.
So welcome to the Age of Algorithms. Welcome to the algorithmic society. Calculations are in the driver’s seat, and they drive us.
Fasten your seatbelts. This ride could be incalculably bumpy.
Danagram
:] for The Progress Report, a magazine that doesn't exist yet
“Calculations are in the driver’s seat, and they drive us.”
Many ages ago, as a graduate student in sociology, I enjoyed referring to quantitative social research as the “dance of the variables.” We apprentice researchers were then learning to work with what are now called 'metrics,' or measurements of variables, relating to class, ethnicity, gender, education, religion, and the like. We were learning to feed these data through various statistical procedures to produce precise-looking correlation coefficients, regression lines, and significance levels. The resulting tables and graphs looked impressively objective when inscribed by a machine, in dot-matrix characters, onto perforated computer paper.
At the time, it amused me to imagine a researcher racing back from the computer center with the exciting news that “mothers’ religiosity explains an additional 3% of variance in son’s status attainment!” Be still, my heart.
I shouldn’t have been so glib. Quantitative social researchers produce data and analyses that are essential to the survival and sustainability of modern information societies. Their role in the today's world is indispensable. Yet there are dangers lurking in the practice of transmuting lived human experience into remote mathematical abstractions. It is as though society itself were now a Pythagorean construct – a matrix of numbers made flesh, or rather flesh made numbers. It's as if the metrics and models, quantitative representations of social reality, are more real and significant than the lives of the people from whom they derive.
These days, the dance of the variables is evolving in an ominous direction as ‘quants,’ not just in finance but across many other fields, devise increasingly abstract and derivative metrics, models and 'instruments' that drive corporate and governmental policy, derivatives that can be crushingly real in their human consequences.
Call it the dance of the algorithms. Algorithms are those step-by-step calculating procedures that are now making many of our most consequential decisions for us, automatically and robotically, in fields ranging from finance to education, and from political campaigning to sports. (Did anyone happen to see Moneyball?)
Algorithms are like mathematical machines, performing stepwise computations that determine the fates of real people's lives. Here are just a few of the algorithms that now seem to be steering our destinies in the early 21st century:
Wall street financial instruments, including financial derivatives, that can behave in unexpected ways as they collide with other variables in the real economy (not to be confused with economists' models of the real economy, which, weirdly enough, have themselves become variables in the real economy);
Googlish search algorithms, the secret proprietary formulas that determine which search engine optimization (SEO) strategies will fail or succeed, and thus which memes or ideas will die or live on to reproduce;
School funding formulas, and especially those based derivatively on high-stakes testing, that determine which kids will receive adequate educational resources and which won’t;
Political algorithms used by campaign strategists to separate the signal from the noise in economic and demographic data, and thus to calculate which messages to send to whom, when, and by what channels in order to maximize fundraising and voter participation on behalf of this or that political commodity. (See also market research.)
Exam problem: Add two fresh examples of your own.
Not only do programmed algorithms make many of our decisions for us, but they also interact (‘dance’) with each other in unexpected ways that far surpass our capacity to predict or restrain their interactive effects.
The derivatives that Wall Street mathematicians concocted prior to capitalism’s last near-fatal crash in 2007-8 played out in ways that even their makers failed to anticipate. Other algorithms across society are now propelling us in uncharted directions that no one fully fathoms. Like all social inventions, they can have unintended consequences for good or ill.
Algorithms have now joined the dance of the variables, becoming social factors unto themselves. Mathematical procedures are coming to replace conscious human decision-making with automated decision-making in a widening array of social institutions.
I’m not suggesting that all algorithms are malignant. On the contrary, some may do immense good. I, for one, would be happy to see a politically neutral algorithm (if such were possible) designed to draw fairer legislative redistricting lines, to replace the current DeLay System, which operates intentionally and effectively to disenfranchise millions of Americans.
Algorithms can be helpful tools. Yet they may also inadvertently create mathematical disasters equaling or exceeding the destructive force of natural disasters such as hurricanes. We’ve already seen the effects of one such disaster in the past decade, the role of quants and derivatives in the financial cataclysm of 2007-8, responsible in some part for the near collapse of the global economy.
Now comes the really scary part. Our multiple algorithms may interact with each other, directly or indirectly, to produce inadvertant Superalgorithms (or more accurately, networks of discrete algorithms) whose potential combined effects are profoundly beyond our capacity to predict or control, like mathematical frankensteins, or viruses escaping the lab and hybridizing in the wild. Imagine, for instance, how stock market algorithms might interact with taxation algorithms, affecting revenues and expenditures in algorithm-driven social service and educational programs.
Even the smartest people in the room (academic, governmental, military or corporate) may be unable to save us from the consequences of these accidental interactions among mathematical machines. Indeed, the smartest people in the room may well be the very ones who unleash them into the world. And like the sorcerer's apprentice, they may find that the forces they unleash have an inhuman life of their own.
So welcome to the Age of Algorithms. Welcome to the algorithmic society. Calculations are in the driver’s seat, and they drive us.
Fasten your seatbelts. This ride could be incalculably bumpy.
Danagram
:] for The Progress Report, a magazine that doesn't exist yet
No comments:
Post a Comment