Posted on 03/21/2017 10:13:46 PM PDT by nickcarraway
Your life is dominated by algorithms and you know next-to-nothing about how they work or what consequences they have.
With Facebook's news feed controlling the (fake!) news we see and the algorithmic robo-debt debacle engulfing Centrelink, it's time we knew more about the algorithms having a growing impact on our lives.
Has an algorithm affected your life? Do you work with any algorithms we should investigate? Tell us about it.
Key points:
We're looking into how algorithms are changing our lives It's more interesting, important and complicated than you might imagine We've collected some interesting reporting on how algorithms are influencing us and affecting our lives We want your help: are there any algorithms you think we should investigate? Get in touch The intrusion of algorithms into our daily lives is much more pervasive than these examples suggest. So pervasive, in fact, that it's easy to overlook their use almost entirely or dismiss it simply as a natural consequence of modernity.
While there's a growing body of reporting on the subject, we know precious little about the algorithms which government and business routinely use to make important decisions about our lives.
There are at least 20 separate parts of Australian law that allow the government to give computers the power to make decisions. Decisions that used to be made by a human and can have important consequences.
These laws allow for computers to make decisions about social security, taxation, parental leave, superannuation, migration, bio security and child support. In every case, some kind of algorithm may be used to make decisions, yet we have no knowledge of how these work.
Wait, what is an algorithm?
In a broad sense, any set of instructions for completing a task or solving a problem could be considered an algorithm — directions provided to someone on the street for finding a nearby attraction, for example — however more commonly we think about algorithms as something to do with a computer. Computers are used to automate the carrying out of instructions defined in an algorithm.
It's this automated, computer-based use of algorithms where data and code drive systems for solving problems and the scale and speed with which they're being adopted that has experts concerned. Not only are they being adopted quickly, but they can make decisions quickly too, so when something goes wrong, it has the potential to go wrong in a big way.
But through all of this it's worth remembering that the errors and biases of computers are human errors and biases. In one way or another, we're still in control. We tell the computers what to do, via algorithms.
A recent report by Pew Research Centre sums up the concern of experts who study their effects succinctly, stating algorithms can "put too much control in the hands of corporations and governments, perpetuate bias, create filter bubbles, cut choices, creativity and serendipity, and could result in greater unemployment".
Algorithms are also getting far more complex. There is a new breed of algorithm, commonly referred to as machine-learning or artificial intelligence (AI), for which it's impossible to know exactly how they work.
This makes it all the more difficult to scrutinise their effects. And we're on the cusp of seeing these new "black box" algorithms take centre stage. The technology behind them is becoming more accessible every day; companies like Google, IBM and Amazon are releasing source code and commoditising AI.
What are the algorithms that affect your life?
Over the coming months we plan to explore algorithms and take a closer look some of the many that affect our lives in hidden ways. We plan to investigate specific algorithms, the systems that use them and the (sometimes unintended) effects they have.
Along the way we'll try to help you understand:
What is an algorithm and what different kinds are there? What kinds of decisions do algorithms make? How close are we to a future that looks like an episode of Black Mirror? Collected below are a few interesting examples of reporting on algorithms and their consequences.
Do you know of an algorithm you think we should investigate? Contact Simon Elvery.
The new taste-makers
One of the first arenas in which algorithms started to have real impact in taste-making was music. Ben Popper wrote about algorithmic music curation for The Verge in 2015. It's still an insightful and interesting read:
"That first track was a risky selection, and the rest of the playlist was, too. It felt like an intimate gift from someone who knew my tastes inside and out, and wasn't afraid to throw me a curveball. But the mix didn't come from a friend — it came from an algorithm." Another interesting and more recent piece exploring similar topics is How Has Streaming Affected our Identities as Music Collectors? by Cherie Hu.
Reinforcing structural discrimination
The Washington Post reported in 2016 on evidence that Uber appeared to be offering better service in areas of Washington DC with more white people:
"While any sort of racially biased agenda by Uber is extremely unlikely, our results suggest that race does play a role in predicting the service quality of uberX in different neighbourhoods. This raises all kinds of substantive questions." Who should get parole?
The use of algorithms to predict whether prisoners will re-offend, impacting on whether they should get parole, is in wide use in the US justice system. There's been a lot of deep and significant reporting on this, much of which has been well summarised by reporters at the Washington Post:
"What does it mean for an algorithm to be fair? Surprisingly, there is a mathematical limit to how fair any algorithm — or human decision-maker — can ever be." Should we work with or attempt to subvert algorithms?
It is, of course, not very new to attempt to trick an algorithm. Indeed, a whole industry — search engine optimisation — was built around gaming Google's famous search algorithm. But what about trying to subvert (or even work with) algorithms for good?
"To curb the spread of unreliable news, should news organisations and platforms turn to algorithms or rely on users themselves? One recent experiment suggests a potential solution could combine the two."
You expect them to tell you?
It is uncomfortable to read an article when every reader of the article is certain to know more than the author about the subject. Computer software always has algorithms. It is not a skill they have just acquired. And the use of algorithms goes back before computers existed.
I don’t like where this article seems to be going. Can an algorithm tell me where it’s going or can it only tell me where it thinks it’s going?...The latter, I think.
FAKE NEWS!
3185296328795229756118816045512425452217437024522750010755991887789210262000017594820131026302153318
If that is true, I wonder how many tickets she had to buy to figure out the algorithm.
I’m thinking that well-designed algorithms could eliminate bias. For instance, an algorithm that screens college applications and weighs relevant grades and other performance factors in order to determine whether applicants are suitable and likely to finish college.
Not that such an algorithm would likely be used for long. The race-baiters have too much vested interested in accepting low-achieving students who are not white or Asian to allow any unbiased selection method to survive.
This shows the authors total ignorance. For most of us who are educated, algorithms pre-date electronic computers and go back to the ancient Greeks.
When it bleeds, it leads is a famous news selection algorithm.
First we identify a problem. Then we assume it is the role of government to solve all problems. Then we discuss which government program considers all the factors of fairness in addressing the problem. Then we discuss behind the scenes how to present the best government program to the public in a way that the public will accept it. Then we present it to the public. Then, when the public resists, we attack ad hominem the leaders of the resistance and destroy their credibility. This is another common algorithm in the news media.
Computers doing algorithms are necessary solely because the volume of data is too large, the the number of problems for the government to solve is too large, for humans to consider all the points of fairness and unfairness.
Big government didn't work in the past (eg LBJ's best and brightest technocrats) because not even the geniuses could handle the massive amount of data. But now with Big Data in Big Computers we can finally get Big Government right.
How's that for an algorithm?
Next to ever lottery vendor is a waste basked where losers toss all their losing tickets. It is quite easy to collect many thousands of tickets. If they have a bar code the computer can easily narrow down the winning combinations to a small number. You then play each of those combinations.
Buy 20 tickets and win $1,000 in a scratch off.
She bought thousands of tickets, IIRC.
She was able to deduce which stores would get the winning scratch-off lottery ticket. Although the store selection was ostensibly randomized, she was able to identify a recurrent pattern that was reliably predictive. She still had to buy thousands of tickets from the selected store to ensure that she had a winning ticket.
...for which it’s impossible to know exactly how they work.
It’s dumbfounding to learn that the cyclical behavior patterns of a former VP of the United States can have such an effect on my life.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.