Big Data Practices Contribute to Inequality

[KentState]

While I agree with most of what you said, this is not really a good statement. Very often the people that the data is collected from is not the customer for the company that is collecting the data. I am not Google's customer, I am it's product.

My argument is that by making your online identity known by using their software, you are in fact a customer. The trade off for the services is them collecting behavioral data. If you however do not use their products, then your identity should not be known to them and you are merely a token. Once again, is Google honoring that relationship? I don't know and the have to self-police and that's where I have issue with big data.

 

[Rebel_L]

I am going to start off by saying that I though the original article did not accomplish showing how data collection creates inequality. The insurance example is especially poor as insurance as a business is strictly about statistical risk assessment. Its the customers wanting a better deal that start the forming of categories to reward the low risk people, the current car insurance premiums already have plenty of discrimination built in, all big data does is to let us discrimination without the biased human factor (or less of it anyways, we do still write the algorithms)

The original article does though bring up the question of data collection and this is where I think we need to make some changes.

I currently work with Big Data for a various large company doing exactly what you describe. We take in hundreds of billions of data points and do build profiles. I wanted to get that out of the way as I do see this from the "other" side.

What I think that most don't realize is that companies want to do this to make money. In order to actually make money, the goal is to align your marketing efforts with your customers. The goal is to not take the data and do something evil and risk the relationship with the customer.

There are a couple points to take away, first I said customer. Meaning that privacy laws makes it very financially painful for a company to mix known with unknown. The companies that do Big Data correctly keep the data in silos and provide a mechanism to opt-out. Second, individuals have to take some responsibility in what they do online or in public spaces. If you openly share every detail about yourself without questioning it, then it can be used to market back against you. Why wouldn't a company want to talk to you in an appealing manor.

On the flip-side, I do feel that there needs to be a serious look at how data is collected, stored and distributed. The trail for input to final output is never clearly defined and most often tied to contract language dealing with a purchase. For example, we sign away so many rights to our data when purchasing a car or something like a cell phone. These are things that so many require in daily lives and it's not as simple as not owning one. Should our personal information become a commodity that we freely give away so we can drive to work?

Right in the beginning is where I start taking issue in how big data works. Companies do indeed do this to make money. Our personal behavior has value and that idea is one that data collectors seem to want to keep a secret. In my opinion data collection and service need to treated as separate items. Everywhere it seems like service is traded for data, yet the concept of data having value is so poorly understood in society in general that it really is an exchange in which only one side understands the contracts they are signing. In some sense this discriminates against the uneducated and that would be the one thing where I believe big data has some impact on promoting inequality.

I personally don't work in or have experience with the fields that would put a value to having my data so I don't even know how much value I give for free on a regular basis. What I think would really be good to see though would be for all companies to have to offer there services for currency, and then in turn pay people for being allowed to track and store their behavior. Like for instance Facebook charges $10 a year to use it, and they pay you $10 per year to be allowed to track you on their site and tailor adds etc. (and what they do with your data should be spelled out in the contract that you sign where you get paid) I think this would have two benefits, one it would show people what value their behavior has, and two it would change data collection from an opt out system to opt in system to help cut down on companies exploiting their customers.

On a side note I think this would help our society have a fundamentally better attitude towards online piracy of other peoples work, as people would start to become familiar with digital data that having real value.

 

[OverVolt]

..

 

[Dessicant]

^^

Oh but it does create inequality. There is a huge hurdle to gain upward mobility. Having a degree even lowers your car insurance rates. Its a double whammy. Without a degree you likely make less, and pay more for car insurance. These little things really rack up. Big data is not on the side of the lower middle class. The more they learn about them it seems the higher their rates go, which kind of defeats the purpose of having a large risk pool for insurance. The risk pools should be larger, not smaller. Big data creates smaller pools of people. More targeted advertising, sure but its also more targeted insurance rates, which kind of defeats the purpose of insurance eh? It turns some years they win, some years you win, into, they just win all the time, every year.

One egregious example is that driving between midnight and 4am in the eyes of the insurance company raises your rates. Why? Correlation, they think you're more likely to be a drunkard. Doesn't matter if you drive at that time for your job. They feel entitled via statistics to charge you higher rates because of the time of day you drive, when its really just a correlation is not causation fallacy, drunky .

Driving at night is more dangerous statistically than driving during the day. It has nothing to do with being a drunk. It has to do with the fact that at night we can't see things as well or as far in advance. Constantly driving under these conditions means you are more likely to have an accident.

And every time an insurance company figures out a new way to charge risky drivers more, those of us who are good drivers can be charged less. I favor smaller and more granular data pools. Risky drivers should be charged more. If that makes it so they can't afford to drive at all, more is the better, making the roads safer for the good people.

 

[Rebel_L]

^^

Oh but it does create inequality. There is a huge hurdle to gain upward mobility. Having a degree even lowers your car insurance rates. Its a double whammy. Without a degree you likely make less, and pay more for car insurance. These little things really rack up. Big data is not on the side of the lower middle class. The more they learn about them it seems the higher their rates go, which kind of defeats the purpose of having a large risk pool for insurance. The risk pools should be larger, not smaller. Big data creates smaller pools of people. More targeted advertising, sure but its also more targeted insurance rates, which kind of defeats the purpose of insurance eh? It turns some years they win, some years you win, into, they just win all the time, every year.

The drive in insurance to make smaller and smaller risk pools has been around long before big data, its customer driven, its because I want to pay as little as possible for my insurance, and so does my neighbor, and my co-workers, and probably even you. Insurance companies have to compete for these people, yet in a sense still have a monopoly because you need to have insurance (at least here you do, I don't know if you can legally drive in the states without it) All the insurance companies want to lure in the most profitable groups and make more money and so fight over them... the high risk groups get screwed over in the process... after all insurance is a fixed deal, all you can do is shuffle around who pays how much, its not like you can innovate your product or ramp up production. The problem with insurance inequality is taking a wealth redistribution program and turning it into a for profit business. Does big data make discrimination more efficient in insurance? YES, is it the cause or the driving force behind the discrimination? NO

Perhaps we should rally against science as one of its main features is to categorize things for the most trivial differences, I mean really, is the fair that we value gold so much and lead hardly at all just because of a few subatomic particles?

One egregious example is that driving between midnight and 4am in the eyes of the insurance company raises your rates. Why? Correlation, they think you're more likely to be a drunkard. Doesn't matter if you drive at that time for your job. They feel entitled via statistics to charge you higher rates because of the time of day you drive, when its really just a correlation is not causation fallacy, drunky .

Could it perhaps be that reduced visibility and increased likelihood of fatigue also causes the risk for driving at night to be higher? I don't know where you get this idea that they believe you will drive drunk if you drive at night. I once heard a stat that a disproportionate number of accidents happen within close proximity to ones home, yet no one was trying claim that its because drunks don't get far before they crash.

The idea of having a system in your car that monitors how you drive and then gives you discounts if you drive well is actually the first innovation in the insurance industry that gives you some upwards mobility to move out of your risk group, you can now work nights and not be a drunk, all thanks to big data.

 

[Rebel_L]

Our basic views are at odds, so of course everything built on those views is also at odds. So the mechanics of what is built upon our premises are not worth debating. I believe your basic view is collectivism-utilitarianism. That, quoting Spock, the needs of the many, outweigh the needs of the few, or the one." That the economy is a tool of society, to be run by governments and other policy makers and experts, for the purpose of making the average person or the "little people" comfortable and happy, even if it means curtailing basic freedoms.

My basic view is that the economy is not the property of anyone, but a sum of the economic activity of its participants. Nothing more. I don't believe the needs of the many outweigh the needs of the few. In fact, I don't recognize need as a claim at all. I don't accept that because one, or many, have a need, that they have the right to take what they need from others.

Moving to income inequality. I don't care about it. Your income is not my problem, and my income is not your problem. I am not my brothers keeper, and neither are you. Everyone needs income, and everyone has a right to act rationally in a free society and a free economy to trade time and/or talent for money to survive and enjoy.

If you have little talent and are lazy, poverty will be the logical result. If you are driven and talented, there is potential for great wealth. If you are like most people, in the middle, you will have a good life, but will have to work diligently and make sure you always have something to offer in trade, but probably won't own a yacht. If you are handicapped or useless, you have the right to ask others for help. If everyone says no, you have a right to die in the street.

So, you believe in collectivism-utilitarianism with curtailment of freedom as an acceptable alternative in the interests of the common good. I believe in individualism and freedom and reject the idea of the common good altogether.

These viewpoints are incompatible and cannot be argued without one of us giving up our basic moral convictions, and obviously that's not happening!

So I believe the correct forum behavior would be for us to state our convictions and move on. Which we have now done.

Im curious as to how you would structure a society, government, laws and their enforcement, taxation etc. The rejection of the idea of common good altogether would seem to be a direct attack on the basic premise of structured society. Rational self interest changes from person to person as their goals in life and skills can all vary, with everyone wanting to advocate a system that benefits themselves the most how do you reach a consensus on what system to actually use without some people accepting compromise that benefits others more than themselves. I would even say that taking rational self interest to its conclusion would look like everyone receiving an equal share of the whole pot with everyone being assigned the task they have the most skills in and doing their best to make that pot bigger so they can get more.

 

[OverVolt]

..

 

[Rebel_L]

Its completely wrong, you're rates will go up guaranteed. The risk pool it puts you into is so small (could even be as small as you are the only person who drives the way you drive and participated in the program) that the statistical unknown factor is so large that your rates WILL go up. Its statistics. Big data cracked the problem, hooray! A risk pool of just 1. Oh wait shit, now the error bars are huge. Lets just charge alot to cover our bases?

And all the questioning over is it the visibility or presumed tiredness? Is it that the bars close at 2am or is it people outdriving their headlights? Kind of shows the problem with correlation is not causation. They have no problem charging you more for a correlation even if it has nothing to do with the cause. If you string enough correlations together, everything is bad for you. Just look at all the correctional health studies. Eggs are bad for you. Eggs are correlated with smoking, eating bacon, and eat lots of red meat. But adjusted for those factors, egg are actually good for heart health. Who knows. Red meat is bad for you but lots of steak in the Mediterranean diet is good for you. Its probably because the average persons "red meat" comes from mcdonalds and stuff where as when you say "red meat" most people are thinking of a big juicy steak you made at home from wegmans, which is probably the type of red meat per the Mediterranean diet that is good for you. So you could cook a ton of steak at home and because everyone else is getting their red meat from mcdonalds they could see that as a risk, I suppose metaphorically per car insurance rates. Insurance companies and big data have gone into a correlation frenzy and ultimately the data is a useless pile of shit because correlation is not causation and they've taken it way too far.

If you are unhappy about how some people interpret data that is another matter all together unless your position is that good data interpretation is not possible. Car accident stats and insurance payouts that accompany them is a much simpler equation than human health and I would guess insurance companies have the risk factors nailed pretty well. And yes poorer visibility and greater sleep deprivation have a huge impact on driving safety. You sound you like don't ever drive in the dark and have never worked night shifts when you say that those are not factors in driving safety. I however don't want to spend lots of time arguing about data interpretation and correctly assessing risk factors because that is not part of this discussion in general, unless your position is that good data interpretation is not possible. People who interpret the data better should see more benefits from doing so and would in turn have a competitive advantage over those who don't, driving the quality of data interpretation to continually improve.

As for this insurance company device the description I have seen is that it is a way to get a discount from your current rates. It being voluntary means you can opt back out and return to regular rates when you choose so I don't see how your price ever goes up, it really sounds more like a way to get the average person to pick up better overall driving habits by incentivising it. If that is not how this particular device from whatever insurance company operates than we are talking about two completely different things.