Welcome, Instapundit readers…and everyone check this out as well…
Let me pull something over from my professional life for a second, because I think it’s consequential enough that you folks ought to know about it. It’s not something I’ve done but something I’ve been reading about.
Google and other companies (Six Apart, among them) are going to open their API for social graphs.
The short version: Google will announce a new set of APIs on November 5 that will allow developers to leverage Google’s social graph data. They’ll start with Orkut and iGoogle (Google’s personalized home page), and expand from there to include Gmail, Google Talk and other Google services over time.
What’s a ‘social graph’, you ask…it’s a map of the connections between people and between people and content.These graphs are deeply meaningful because they are not random; they are full of meaning because they track our behavior, contacts and interests.
When I was designing my version of Pajamas Media, I wanted to construct a social graph connecting readers and blogs and use that data to predict what new blogs people might like, and to cluster ad placements to targeted groups of readers. I thought – and still think – there’s a ridiculous amount of value in that data. There is also a series of problems with that data, and they are at root problems of privacy.
Because the tradeoff for the usefulness of having someone suggest new blogs I’d find interesting is that someone has to know what blogs I read. And to the extent that I read blogs that I don’t want people to know about – blogs about sex, psychological issues, political positions antithetical to my public persona – that’s potentially a problem.
I thought I had an elegant solution to the problem for my PJ’s design (hey, I can’t tell all my secrets), but still saw (and see) this as the crux issue a system like this will have to get around.
But now much bigger fish are stepping into the market, and they are doing it with data much more serious and personal than your blog-surfing habits.
Google is announcing that it will create a series of open API’s (data interfaces) that will allow other people to write systems that will allow them access to an undetermined set of Google’s social graph data. What data does Google have? Well, pretty much everything. My email from my Gmail accounts; my search history; my blog posting and links if I keep them on Google tools; the links to and from my blog if the Google spider is set to capture them; the YouTube videos I watch, and so on…
Now the folks doing this are serious and smart people. They have explicitly talked about the issues of privacy. here’s Six Apart’s David Recordon:
An open social graph is just as important as an open identity.
* You should own your social graph
* Privacy must be done right by placing control in your hands
* It is good to be able to find out what is already public about you on the Internet
* Everyone has many social graphs, and they shouldn’t always be connected
* Open technologies are the best way to solve these problems
* We’re going to release code and demos soon
The privacy and security implications of this are pretty staggering, if I’m interpreting this correctly. They are solvable – I did a baby solution as noted when I designed the system for PJ’s, and I don’t doubt that the horsepower of this group can propose useful solutions.
But I’d be a helluva lot happier of they had started with the basic principles and mechanisms for ensuring privacy and announced those first – before releasing working code modules.
I’ll be digging into this more deeply, and may have some posts here to talk about it. Meanwhile, discuss among yourselves, and personally note that I welcome our new Redwood City overlords…
It may be that Big Brother is here and he is watching. Or not, depends. Tell us more when you have time. Personally I always do what I can to secure my identity – firewalls, pseudonyms, never ever use that damn wallet or autofill, etc.
I have purchased lots and lots of things via the web in the past years and have yet to have my identity stolen, but that, as it may be, only a matter of time.
I love Google tools BUT I may have to abandon them should they become too intrusive.
I must get that TOR proxy stuff up and running as soon as possible…
And remember that TOR has recently been implicated in man-in-the-middle sniffing and data mining. Encrypt all you like, there remains traffic pattern analysis. Do you trust *every* jump in the mixmaster?
Then there’s the report that a majority of the TOR user base doesn’t even bother to encrypt.
#1 from Robohobo at 6:02 am on Sep 22, 2007
It may be that Big Brother is here and he is watching. Or not, depends. Tell us more when you have time. Personally I always do what I can to secure my identity – firewalls, pseudonyms, never ever use that damn wallet or autofill, etc.
__________________________________
It is disconcerting to me that we have to resort to this much work to secure our privacy. Especially, since we seem to be fighting a rear guard action as our right to privacy is constantly being forced into retreat.
If there was anything that was ripe for a pre-emptive strike it would be the information gathering industry. Perhaps an bill of electronic Privacy rights should be passed into law that would define what you can and cannot collect on the internet.
I am well aware that their is essentially no privacy on the net. I would like to see some efforts directed at trying to reclaim some.
I am a big fan of Google, but am increasingly concerned about their continually offering what amounts to Trojan Horses to their customers.
A few years ago, a perspicacious friend of mine suggested that the battle was already lost: personal privacy was gone and wouldn’t be coming back, and moreover, that young people (even if they didn’t think about it) accepted and embraced that. His examples were social networking sites, where information that us older folks would consider intensely personal is just given away. It took me quite a while to realize how right he was. Having done so, this development does not overly trouble me – it was a matter of time.
But, and there is a very big “but” to it, there are some serious implications to be worked through socially. Can we expect our employees to be saints, when it is possible to discover everyone’s sins and when everyone has sins? What about our public servants? And if everyone’s misdeeds, hidden quirks and peccadilloes are to be publicly accepted, because otherwise everyone would be held in public contempt by large numbers of people, then how do we prevent the norms of good behavior from inexorably sliding towards public perversion and decrepitude? The problem with appeals to the better angels of our nature is that about 10% of us are sociopaths and frankly don’t have better angels, but are largely only kept on good behavior by standards of public decorum and fear of ostracism if they violate those standards.
It’s gong to be an interesting world.
Let me connect this with another can of worms: insurance.
The theory behind insurance is very simple: spread large but rare risks across a statistical universe, so the average exposure become quite small. An insurance company can make a reasonable profit while protecting the individual customer from disaster.
But the key term here is “statistical universe”. The larger the statistical universe, the closer the customer’s premiums can approximate the average risk.
However, there is an obvious incentive for companies to maximize profits by restricting their risk pool to the lower-risk part of the overall population. By reducing overall payout, they can decrease premiums to low-risk customers while increasing their profit margin. (That is, even the low-risk customers don’t get the full benefit of the savings.) And, of course, the high-risk customers are in bad shape, with very high premiums or no coverage at all.
What does this have to do with data mining? Everything.
The more data available for mining, the better companies can make fine-grained estimates of individual risk. If insurance companies are allowed to segregrate their risk pool based on those estimates, then genuine insurance becomes largely unavailable. If you are basically healthy, you can ensure against truly random accidents. But if you do have risk factors, you are out of luck.
The information genie is out of the bottle. There is no way to prevent access to all this information, that makes all sorts of risk estimates (and other privacy invasions) possible.
It is, however, possible to regulate the nature of the risk pools that a company is allowed to create, in order to make a small number of very large statistical universes, for risks to be averaged over.
We need to ask ourselves whether, as a society, we are better off by allowing people to sicken and die without health coverage. Just as your kids are better off if all the kids in their classrooms are inoculated against infectious diseases, I believe that even those of us who are currently healthy are better off if everyone who is sick gets good health care.
When you look at data mining, it is important to think about the huge range of decisions that might be made, using that data. This is a key one.
Suppose that instead of for-profit insurance, we let the government pay for everybody’s health care as best it can.
Then data mining to make insurance more profitable becomes irrelevant.
What might then be useful is to make aggregate medical records widely available. Do enough obfuscating that it takes an effort to find individual patients. Make it illegal to pay attention to individual patients. Work to establish a social climate that says it’s disgusting to do that — because law without social climate behind it is pretty flimsy.
Then anybody could look at data about competing medical treatments. Anybody who wanted to devote the time to it could use the actual data to guess which treatments work better. I think that might be a good thing.
For example, suppose that some psychological maladies have a slow cycle of months. A patient takes a drug for such a malady and either by placebo effect or coincidence then feels better. They keep feeling better for a few months and then feel worse, and get their dosage upped. Then a few months later they feel worse again, and again get their dosage increased. Eventually they get to dangerous levels of the drug and their physician switches to another drug, and they pehaps improve and perhaps not. They switch drugs every month or two until they improve, and then the cycle starts over.
At first sight it seems strange to put the obvious improvements onto placebo effect, but placebo can be very powerful. Currently it would be hard to test this idea, but if the data was available it might be easy to make it at least look plausible or implausible. And if the expensive toxic drugs turn out to be unneeded, that might reduce costs. Or if another expensive treatment replaced those drugs, at least it might be an effective one.
There are lots of medical improvements we might make if the data was easily available. For example, a whole lot of surgery is done on patients who don’t particularly have any viral or bacterial infections when they first go to the hospital. But a lot of sick people wind up in hospitals and nosocomial infections are common. Why should infected patients go to the same hospitals as uninfected patients? Tradition. And the numbers aren’t obvious to the public, although everybody knows it happens.
“link”:http://www.cdc.gov/ncidod/dhqp/nnis.html
The experts try to monitor it a little bit, and report their results. Not obvious. If it was clear about the costs of current therapies we could much better compare the costs of splitting the hospitals.
J Thomas — Experience in Britain, the other EU states, etc. shows that Government paying for health insurance has no effect whatsoever, other than to shift coverage decisions to whatever political group has the most political power. Thus in Britain, middle class people are denied surgery for broken ankles while expensive reproductive treatments (IVF) for gay couples are a basic “right.”
All putting Government in the mix does is make the fight a naked patronage fight. A dollar denied a middle class straight white man for example not only punishes that man (a positive good for Liberals regardless of country) but is available for politically correct identity groups.
I find Liberals concerns for privacy laughable. Google is “good” i.e. it does nothing on it’s home page for 9/11, Memorial Day, etc. but celebrates Roald Dahl’s birthday (a notorious anti-semite). Thus any and all privacy intrusions by Google is viewed as “good.” Why not? Isn’t Google on the right side. Yet the NSA listening in on terrorist conversations is the cause for the vapors.
To my mind, the real risk is the unregulated, totally without oversight, privacy intrusions of Google which has a habit of turning over data to China. NOT the NSA which lacks the people, money, power, and know-how to turn over my private data to the Chinese for money. But Google’s anti-American politics are “correct” so they get a pass from the Press and Liberals.
“Suppose that instead of for-profit insurance, we let the government pay for everybody’s health care as best it can.
Then data mining to make insurance more profitable becomes irrelevant.”
Have you seen the story from Britain where the NHS is refusing to treat a man’s broken ankle because he’s a smoker, he won’t quit, and supposedly that makes the treatment less likely to work?
The government has exactly the same motivation to ration health care as the insurance industry (collect as much as it can in taxes, pay out as little in benefits, use the overage to pay for something else, whether a house in the Caymans or fund light rail), plus whatever ideological causes are hot today. I’ve been told that because I’m not a liberal, I’m mentally ill. Wonder how long it would take Hilary or Kos to put me on the “do not treat” list? Buy a steak? no bypass for you unless your diet is PETA approved.
The government has no incentive to data-mine? What color is the sky on your planet?
No, J Thomas, what data mining would do for a government with absolute control of health care services is allow it to be more efficient at stealing babies.
Follow the link; that isn’t hyperbole.
[#8,9,10] Sounds like talking points harvested from Coulter and Limbaugh, rather than a rational argument. No citations I notice, except for one irrelevant anecdote by Mark Poling. The situation about out-of-control social workers taking away babies in the name of preventing child abuse is certainly appalling, but it’s not relevant to how the society chooses to organize its health insurance.
Jim Rockford: you have made “Liberals” into a boogieman that seems to scare you, but doesn’t resemble anything I am familiar with. You make claims that should be backed up with statistics, not anecdotes, but you don’t have citations for either one. I imagine that there is a horror story about someone with a broken ankle, and another one about IVF. There are also horror stories about people in the US being denied healthcare for bogus reasons by their insurance companies. A pissing match between horror stories isn’t rational discourse. Rational discourse takes a bigger picture, which typically requires statistics.
Sorry to take the fun out of it, but that’s how the big boys play.
I don’t know who told SDN that he was mentally ill, or why, but I’ll bet there is more to the backstory than that he is not a liberal. Lots of people are not liberal, and not mentally ill. What makes SDN special? Perhaps it’s paranoid fantasies about persecution by Hillary or Kos.
The argument about insurance, and the size of the risk pool (or statistical universe), is simple mathematics, combined with elementary capitalism. If you don’t regulate the composition of the risk pools, all the incentives are for the insurance companies to off-load the risky parts of the pool. If you don’t want your society filled with sick people with no access to health care (a serious public health risk), you pretty much have to prevent that natural profit-maximization behavior by insurance companies. It’s about as straight-forward as saying that people have to drive on the right side of the street, or stop at red lights. Yes, both of those infringe on your and my freedoms, but society as a whole work better if we all obey those rules. Deal with it.
Beard and JT, I work for a major insurer in the health care sector. We do exactly the type of data mining you are describing (we use a suite of SAS applications).
What we often find is large (and costly) variance in the way providers’ practice patterns – meaning that controlling for demograhpics and diagnosis and severity, providers will provide a wide range of treatments.
We look at he costs and benefits and attempt to assist providers with establishing best practices. We also, in some cases, use the evidence from the data mining to write medical (meaning what we cover and what will not cover) policy.
It is my experience that it is not the insurance companies that are causing high premiums through profit seeking or even excessive admin costs (the company I work is not for profit), but it is profit seeking providers that are driving costs up. Furthermore, patients are not capable of making economic decsions when it comes to health care. This is understandable to a point. The marginal cost of extra consumtion to the insured patient is $0. So even if an expensive procedure has litle chance of improving the health of a patient, the patient is still willing to give it a try. Patients have no incentive to match marginal cost to marginal benefits. Nor does the provider; there is a principle/agent problem in that relationship. Only an at risk insurer has incentive to control cost and the infrastructure to do so.
The major drivers of health insurance cost are 1. provider practice variation 2. the growing range of new technologies that make it possible to treat more people and for a greater scope of conditions. 3. Expensive new technologies that enhance existing products/services, but do so at a cost that exceeds any extra benefit derived. 4. An aging population. 5. Fat sedentary people who allow themselves to become subject to acute and chronic conditions due to their life styles.
Universal health insurance, single payer, socialized medicine will do nothing to address any of these problems, so far as I can see.
Finally, a large number of the uninsured are young and healthy people between ages 18 – 30. Actuarially speaking, this population comes out ahead by not having insurance. Their avg annual health care costs are less than the the cost of insurance. Even in the event of serious costly illness they tend to lack the assets that would prevent them from becoming Medicaid eligible; this despite horror anecdotes to the contrary.
Yes, putting this group into the would lower the avg cost of insurance, but it would be a tax on this population; a wealth transfer from young people just starting in life to older sick people. THis aspect must be recognized. Also, it is a one time fix that will not divert the long term trend toward higher costs for the reasons I enumerated above.
My $0.02, anyhow. But I do know what I’m talking about here.
If we’re going to do accual insurance, then is there really anything wrong with datamining?
The theory about insurance is that you pay in because you don’t know whether you’ll need to collect. But if you do know, why pay in? It makes perfect sense to have low-rate insurance for nonsmokers and high-rate insurance for smokers, who have a known health risk. Similarly, low rates for people of normal weight and high rates for obese people.
Once you find out that someone is at high risk for an expensive disease, doesn’t it make sense to bump his premiums to an appropriate level? And if he can’t pay that’s unfortunate for him, but why should the low-risk people carry him?
To argue that insurance shouldn’t be allowed to use their knowledge depends on some sort of social utility theory. We don’t want sick people to run out of money? Like we don’t want poor people to run out of money? Hey, in theory insurance is about pooling risks, not about providing health care to people who can’t afford it.
But in practice insurance is not just about pooling risk, either. Insurers have a lot of bargaining power. Individual sick people have none. So it makes sense for insurers to bargain down their payments. “Here’s my offer, take it or leave it. If you don’t take it another hospital will.” And with payments for the insured below cost (though of course not below variable cost) who makes up the difference? The uninsured, of course!
If you’re uninsured and you pay $90 for a visit to the doctor, do you think insurance companies pay enough after the co-pay to bring it up to $90? No way. They pay less, and you are paying more to make up for it. If there was no insurance your uninsured costs would be lower. They help to *make* insurance absolutely necessary. Insurance companies make it *more* expensive for the uninsured to get routine care.
A big part of what insurance does for you is collective bargaining. We usually think that’s bad when it’s unions, why is it good when it’s customers instead of employees?
More later, this is getting too long.
“The situation about out-of-control social workers taking away babies in the name of preventing child abuse is certainly appalling, but it’s not relevant to how the society chooses to organize its health insurance.”
In what way is having every bit of data about past activities irrelevant to the question of where state power ends? Or corporate power, for that matter?
The fact is that the data will be available, and the fact is I would rather society be organized so government was a check on abusive corporate power rather
than simply a (legally sanctioned and heavily armed) replacement for it.
I can name a couple of societies that bought heavily into the idea that such trates “infringe on your and my freedoms, but society as a whole work better if we all obey those rules”, but I personally would not want to have lived in any of them.
“The theory about insurance is that you pay in because you don’t know whether you’ll need to collect.”
Yes but…..in the health care sector people purchase insurance for another reason as well; to gain access to goods/services they would not otherwise be able to afford. No other insurance market is like this.
True, we insure cars, homes, etc against loss that we cannot afford to experience, but we could originally afford the insured object. E.G. I can afford a $250K home. $250K is the limit of my exposure. I can’t insure the home for $500k. That would be moral hazard. The insurer can price the premium accurately based on risk factors and, most importantly, max exposure.
You can’t do that in health care. Again, the constant introduction of new and costly treatments and ancillary technologies for a widening range of conditions precludes pricing accurately.
Furthermore, possession of certain risk factors would cause the very people that need insurance the most to be priced right out of the market.
A small % – anywhere from 1% to 5% depending on product – of the risk pool typically represents more than 50% of the cost in a year. So yeah, most of us could save a lot if insurance was based on individual risk factors and experience, but, again, the ones needing insurance the most would be facing like $50k/year in premiums. Would that fly????
When I was in grad school, one of the other guys had been an insurance adjuster. He had a college degree, and he managed a roomful of girls with high school degrees in a windowless building in chicago. The company had a set of directives about which treatments were OK, and the girls looked at medical records from all over the country and decided which treatments the company would pay for.
He said they didn’t have a quota system that said they were supposed to disapprove some %. But later he did admit that if a particular girl or the whole floor disallowed too few they got investigated because that was a sign they might be examining the records too carelessly.
Each company maintained their own set of directives. Of course this was duplication of effort, and it would be cheaper to have one set for all the companies, or (in the short run, ignoring the problems that come without competition) one company. On the other hand, with multiple alternative directives different people can get different treatment. On the third hand, nobody gets to look at the directives. You wouldn’t really know what you were getting if you wanted to comparison-shop even if you could see them, but you can’t.
Insurance companies are currently our brake on medical spending. Patients don’t care about costs — they have insurance. Doctors don’t think about costs when the patients can pay. It’s only the insurance companies that care, and they show their care by telling your doctor he can’t give you the treatment he thinks you need. There’s nothing you can do about it after you’re sick, you sure can’t switch insurance companies at that point. Probably you got your insurance through your employer, who perhaps offered you 2 or 3 choices, and you couldn’t figure out which one was best.
Suppose the aggregate medical records were available to everybody online. Not so available that everybody could easily look up your STD treatments with your name on them, but available enough for anybody who knew how to do statistics on them.
Say you or one of your relatives has a problem, and you look up treatments. You find that one treatment has a 6-month survival rate of 10% while another has a 6-month survival rate of 20%. That gives you the chance to settle your affairs and say goodbye and all that. If the better treatment is more expensive you might petition to get it. Or you might look at the side effects and figure you’d rather just take your chances and die. How many people to give each treatment to would be a little bit of a political issue, as happens when citizens discover the issues. But would you really want either treatment? Why not petition to get into an experimental group where you *don’t know* the odds?
Say you find out about two treatments and one is much more expensive than the other but there’s no clear advantage to it. Would you want it? If all you know is that it costs more, you might. It’s only natural to suppose the more expensive version is better. But would you make that mistake when you have data?
When people are trying to use political connections to get more expensive treatment, it seems to me that a very good response is “We are running a nation-wide experiment. You were randomly assigned to the control group. If you want out of the experiment we can give you the older established treatment that we are dissatisfied with.” Would people accept that? I would.
In a reasonable country this ought to work fine. But I have some doubts about the USA. Would anti-abortion fanatics look at abortion data to find ways to sabotage abortions? Would anti-smokers use data on tobacco health risks to get tobacco classed as a dangerous addictive drug? (It is a dangerous addictive drug, but would treating it like marijuana or cocaine give us better results than we get with other drugs?) I don’t see anything wrong with collecting data about lifestyle effects on health and using it to encourage people to voluntarily get healthier. But would americans try to force us to get healthier to reduce government costs? Forced exercise? Forced broccoli? I’m not sure we’d be above such things. It might be an interesting experiment, provided I got to be in the control group.
_Yes but…..in the health care sector people purchase insurance for another reason as well; to gain access to goods/services they would not otherwise be able to afford. No other insurance market is like this._
Yes. This is the biggest part of why we have a crisis, isn’t it?
If it was 1% of the population that needed catastrophic care and 99% that needed nothing, people could pay into it like a lottery. A lottery you hope you don’t win.
But what if it’s 50% of the population needing catastrophic care they can’t afford, and 70% of the population needing routine care they can’t afford. That isn’t insurance, that’s something else.
So, insurance costs rose until most individuals couldn’t pay them. Employers paid for insurance instead. They had more money, and they could write it off as a business expense.
But now insurance costs are rising to the point that many employers can’t pay them. So they hire more of their employees part-time or contract so they don’t have to insure them. And they want the government to pay the insurance costs they can’t afford.
And when the government can’t afford it either….
_Again, the constant introduction of new and costly treatments and ancillary technologies for a widening range of conditions precludes pricing accurately._
That’s free enterprise, man. The medical technology companies get big profits once they can get a costly new treatment approved by the government. And the insurance guys can’t just refuse to pay for it — they need evidence it isn’t worth doing, and by the time the evidence is available the technology will be improved or replaced by something newer.
If the government was central to that, they could budget money for R&D for new med technology, and there would be a stage where the stuff gets tested on a randomly-selected fraction of the population it’s supposed to be good for. They could estimate the costs and keep deciding from the beginning how much of an improvement it would need to make to be worth the cost. We could have a significant fraction of R&D devoted to new methods that are less expensive but just as good. That doesn’t excite venture capitalists so much — the profits have a ceiling even when they win.
_A small – anywhere from 1 to 5% depending on product – of the risk pool typically represents more than 50% of the cost in a year. So yeah, most of us could save a lot if insurance was based on individual risk factors and experience, but, again, the ones needing insurance the most would be facing like $50k/year in premiums. Would that fly????_
That’s a political question, isn’t it? But what if we find a way to turn it into individual choice.
If you pay insurance that doesn’t pay your catastrophic needs when they show up, then you don’t have catastrophic health insurance. We have to pay the 1% who need it. But what if you pitched it to customers this way:
bq. We have two plans. Plan A will pay up to $2 million if you need help. Treatment for an unexpected heart problem could do that. Or cancer. But it will only pay that inside one year. If you’re still sick from the same thing after that we’ll renegotiate the contract, and if it looks like you’ll be so expensive you can’t afford the premiums, you’ll do without.
bq. Plan B will pay whatever you need for as long as you live. See the difference? Plan A gives you whatever it takes to get healthy again if that’s in the cards. But if it costs a lot to keep you lingering in a hospital bed for years, it stops. It’s for people who’d rather die than live as invalids. Plan B will drag it out. Here’s the cost for Plan A, and here’s the cost for Plan B. Which do you want?
Once you choose plan A, then when you turn real expensive with no real prospect of getting healthy, you can figure it’s your own choice and you can check into a hospice and enjoy your dying. And if, say, Plan A is $50/month while Plan B is $1000/month (or more) then a whole lot of people will choose what they can actually afford. We don’t have to choose for everybody. We can each choose for ourselves. When it’s choosing for everybody I feel bad telling even 0.25% of the population they have to die now because I don’t want to pay to keep them alive. (But when the government is funding the R&D they don’t have to create expensive technology to keep people barely alive.) I can choose for myself with no guilt at all.
Thanks for the thoughtful response, avedis. If I get the chance to think a bit, I’ll respond in more detail.
A significant point you raise is that medical care must be rationed. Whether it is rationed by denying some people insurance, or by eligibility clerks at the insurance company, or by people implementing a government policy, it will be rationed. We just have to decide which is the best the way to do it, for our society and our sense of fairness.
Guys – can I pull this (interesting & civil!!) discussion on health insurance back onto the topic? If I promise to do a health insurance post this week?
A.L.
#10 from Mark Poling at 1:59 am on Sep 23, 2007
That the debate is even being done at all tells me all I need to know about the future of a society when supposed ‘health care’ is placed in the control of the Nanny State. Stealing babies is the least of the things they will do. How them re-education camps coming?
Marc Canter (cofounder of Macromind, long ago) has some interesting thoughts in his blog on this topic, here.
Re: that forthcoming health care post–Beard, avedis, J Thomas, and others have raised very insightful points. I say that having just returned from the Genetic Alliance’s “conference on the future of genetic testing.”:http://www.geneticalliance.org/testingsummit We are entering into the age of medically-relevant genetic information (“the thousand-dollar genome”); this will have huge effects on many of the trends discussed here. Consider, for instance, avedis’ notes on risk pools (#12): what if insurers are able to stratify customers on the basis of susceptibility to expensive (i.e. chronic) diseases? Technically, this is already a reality for Alzheimers’ (ApoE allele typing).
I’ll save other comments for the follow-on post, except to say three things.
(1) IT initiatives like the one A.L. discusses in this post may indeed have their largest impacts in the health care field.
(2) One speaker spent years in the U.K., and discussed his first-hand experience with the NHS’s rationing schemes. Come down with ovarian cancer? The budget for Northern Scotland allowed for 10 patients getting (expensive) interferon therapy; the next 30 received only cytotoxic agents. (Make sure you get your diagnosis in January.)
(3) The Genetic Alliance has pushed hard to enshrine genetic nondiscrimination in legislation. This summer saw the passage of “GINA”:http://www.geneticalliance.org/ws_display.asp?filter=policy.leg.nondiscrim by the House; Bush has indicated that he will sign the bill if given the opportunity. Given how November 2008 will capture the attention of Congress next year, “S.358”:http://www.govtrack.us/congress/bill.xpd?bill=s110-358 is probably a bill that will pass the Senate soon, or not at all.
Back on topic, I expect the public will tend to get less scandalised about a lot of things. Some things will be considered *awful* and others will be no big deal.
Like, the way it seems to be shaping up now, if you get seriously suspected of being a child molestor you’ll probably have to leave the country. Anywhere you go, people will find out about you within a couple of days and most of the parents will be outraged that you’re there. They’ll pressure your employer to fire you and your landlord to evict you.
But say you’re a manager and everybody finds out you look at bondage porn online. Then maybe a woman employee might wear a ribbon around her neck and bracelets go into her performance evaluation. Not as an offer, just to look good. No big deal.
I expect a lot of voters will get less concerned about their elected politicians’ personal lives, within reason. Too easy to find out. So homosexuality or attempted wife-swapping etc won’t be particularly good blackmail material. Years ago I listened to some feminists complain about a politician — I think it was Representative Packard, though I’m not sure. He voted for their feminist legislation so they felt like they wanted to support him. He considered himself a feminist. But when lesbian feminists went to lobby him he told off-color jokes and groped them. They were torn between supporting his feminist legislation and opposing him for being a sexist pig. I hope that as people get more open about the things they can’t hide, they’ll care more about the votes.
The government will get more power about the things it cares about. But that’s pretty much inevitable. The more people live in cities the more they have to compromise about living together, and the government handles part of that. The social control that small-town society loses, of course gets picked up by government. But the small towns still exist, and it’s quite possible that small-town people will find themselves thinking about what their neighbors would think every time they click on a link.
JT – and you’re OK with all that?
A.L.
JT, you are much more cavalier about all that than I am.
I am wondering about something. So say someone has an internet account in the man of the household’s name with a static IP address and say his teenage kids and his wife all use the same computer and internet account. He may surf around the net, go out for the evening and then his daughter jumps on board surfs, etc. Then later his son gets on the computer – and god only knows where he is surfing to…..the next day his wife is out on the net using the same IP address………..(this is how it was in my household and I suspect many others)
The profile for this account (the man of the household) would look pretty friggin’ weird since all four family member/users have vastly different interests and reasons for going online. The data would be pretty useless from a marketing standpoint.
But worse, what if the teenage son was, as teenage boys are wont to do, getting into some weird socially unacceptable stuff on line. The activity would be attributed to the father/husband because he is the account holder.
So JT, in your example of the bondage fetish, what if it was the teenage son who was looking at the bondage material and the father is really more of a lace kind of guy?
Or what if someone is doing research – maybe for a doctoral thesis – on taboo topics (e.g. deviant sexuallity, illicit drugs, etc). The research goes on for a couple of years. Sure would look like the person is “into” that sort of thing. All sorts of embarrassing, costly and otherwise truly damaging situations could result from the misinterpretation of the meaning of the data.
Should someone who takes out an internet account be held responsible – in one form another – for all on line activity on that account? Could it ever be proven who actually did the surfing?
I guess it is the possibility of false accusation through circumstancial evidence, innuendo, assumption, etc that bothers me as much as the invasion of privacy.
I would think that there are enough people like me (that just plain don’t like it) that the market will produce new high quality search engines/Google rivals that will not invade privacy in this manner and people (like me) will pay for that service; perhaps even at a premium for the guaruntee that on line activity information is not being gathered, assembled and packaged for distribution.
Avedis,
I agree with you totally. Freedom of the press includes anonymity of readership.
If I should decide, for whatever reason good enough for me (e.g., “know thy enemy”), to spend time reading some truly awful web sites, I should be able to do that privately. I don’t believe it’s right for Google or the Government (without a court-authorized search warrant) or anyone else to snoop into my reading habits.
I think this is more than a consumer choice issue, and being willing to spend more for privacy. I think that Google should have a responsibility, based on their role in the Internet, to guard my privacy, just as AT+T should have a responsibility not to let anyone listen in to my telephone calls without a warrant. (Oh, wait …)
To be honest, I’m rather surprised no one’s yet whipped up a trojan that infiltrates a target computer, downloads lots of child porn and ‘leaks’ that to the FBI.
Or whatever other incriminating payload you wanted. Given the rather shoddy state of computer security, you could damage a lot of people in a hurry doing that.
I don’t believe it’s right for Google or the Government (without a court-authorized search warrant) or anyone else to snoop into my reading habits.
The problem is that it isn’t Google (per se) or the Government doing the snooping. It’s the site you visit that does it. You can argue that 3rd parties have no business with your private dealings, but it’s a lot harder to argue that the people you’re visiting don’t.
When sites let Google put ads on their page, in reality you are then going to Google (to fetch the ad). Now Google knows your IP and where the request came from. And you requested it.
Without online advertising, most of the tracking/privacy concerns on the web fall apart. Without essentially the tacit consent of all the people placing Google ads on their pages, Google wouldn’t be able to put together all this information on people’s browsing habits.
Google employs some of the smartest people on the planet. They should be thinking seriously about how to protect my privacy (and yourse) while collecting and selling all this other information. Or else don’t collect it, and don’t sell it.
If they can’t provide a convincing privacy guarantee, then that aspect of their business should be against the law. I understand that no law can prevent this information from being collected to some extent, but it can prevent it from being a significant part of the business model of a major corporation.
Avedis, I’m fine with society generally getting more tolerant. And I expect it to happen, particularly wrt to online stuff. It’s OK to think about whatever you feel like thinking about. Some things we don’t want people doing.
I think it’s very sad that there’s no way for convicted child molestors to redeem themselves. But there’s no way they can prove they won’t do it again (whatever “it” happened to be) and parents won’t want them around their children. Ever. I read a book from the 1930’s where occasionally somebody who felt too shamed (for falling for a mean practical joke in one case) would pick up and leave town. He’d move somewhere else and start over, with all the disadvantages of not knowing anybody and having to make new friends in a strange place. Because he had no respect where he was. But with the internet you can’t do that. You can get a new ISP and a new name and _blog_ from scratch, but if you move to a new town people can find out what they really want to about you. If it’s something they care about, you can’t hide. You have to join the foreign legion or something.
But the little things will be no big deal socially. So to take an extreme example, say you fantasize about rape. You occasionally read fictional stories about nonconsensual sex, maybe you buy access to porno websites that claim to photograph rapes, etc. Very few people say that rape is OK, but it isn’t a big deal to pretend on the internet. Possibly police who’re looking for a rapist might start with people in their area who have a history of looking on the internet. If they convict some rapists they find that way, they’ll do it more. I don’t have any problem with that provided they aren’t framing their victims. And if that turns out to be an ineffective way to catch rapists then I hope they don’t bother.
We have new technology that provides new power. We can say we don’t like the new power and we want to pass laws to make it illegal. How likely is it for that to work? The more public the action, the easier it is to enforce laws against it. Laws against uttering-and-publishing are quite enforceable. The more people that receive your words the less likely you can get away with it. But things that can be done secretly? Big deal.
I talked with a guy who worked in a bank. He said he could look up anybody’s credit history and anybody’s banking data, with no supervision. Officially people are supposed to have privacy, but bankers sometimes have a legitimate right to know, and who actually regulates their computer systems? They do. If he publishes the data then you can get him easily. If he had sold me my ex-wife’s banking data then possibly I could have blackmailed him over it. A dangerous thing to do, considering.
I talked with a guy who worked for an insurance company. He said he could get anybody’s medical records. People should have privacy and he could get in trouble for revealing their secrets, but he could look all he wanted.
It isn’t new. It’s just extended. More of it. And remember, if Google is forbidden to sell your data, that means that Google can go into business in competition with the guys who need your data, and Google has an advantage that they *aren’t legally allowed* to give up.
When I got a temporary job with Google I had to sign a document promising I wouldn’t look up people’s personal data and use it in any way. I didn’t get any opportunity to do that, but it was part of the standard contract. So if there’s some kind of uproar about Google employees getting caught doing that, Google can fire them for it. Feel better yet? Heh heh.
Similarly with government. We can get laws passed forbidding the government to do various things. And who do we depend on to enforce those laws? The government. Bush has eliminated various protections for whistle-blowers….
We have to live with the consequences of our technology. Like, with nukes we tried for nonproliferation and we got 50 years of minimal proliferation. NPT is dead now and the next 50 years will be different. Learn to live with it. Maybe we can slow down the loss of privacy too. If we get 50 years out of the attempt I’ll be very very surprised.
Marc, I just spoke with several Google employees, admittedly low-level researchers, not the high-level strategic folks. They hadn’t heard of anything like this, and thought it sounded completely contradictory to everything they know about how Google works. By itself, that doesn’t prove it’s false, but it raises the question.
Then I want and looked at the original source, which is a report, supposedly from three out of fifteen high-level people connected with Google, who had signed NDAs and attended this meeting, but who were nonetheless prepared to discuss Google’s strategic plans with a blogger. Something is starting to smell here.
If true, there’s plenty of reasons to be concerned. But I think there are good reasons to suspect that the report might not be true. Do you have independent verification?
Beard – not yet; or rather verification that something is planned but not exactly what.The case set out in the blog post I reference is worst-case (ie. Google chat + Gmail + iGoogle social data, as opposed to just opening Orkut, for example…
Clear as mud?
A.L.
Yeah, but that’s not unusual. Keep us posted.