There already exist flexible materials which can be made to change shape under an applied current, it should be possible to make them small enough to display haptically at the tip of a gloves finger a fingerprint taken by a fingerprint sensor on the inside of the glove.This would solve the problem of the stolen glove as well as the mistaken belief in biometric access control.
You are perfectly right, with people there is no moral decision involved, they just act, most often inadequately in retrospect, and we are fine with that, as strange as it might seem, with rare cases being "tragic", inescapable, impossible to prevent.
This is the problem with algorithms, the accident does not simply happen, it is happening the way it is happening in some explainable way, so people want to know the reasons why it is happening like this and not like something completely different. So it does ot suffice to be better than a human statistically.
Don't get me wrong, I firmly believe we can find acceptable moral algorithms and we should pursue creating them.
This set of rules is one ruleset implementing moral judgement. What needs to be done is finding rulesets people could agree upon, which implement acceptable moral decisions and are complete enough to cover even rare situations in a way people are comfotable with.
You are quite right to require non-discrimination, thanks for providing this insight, as I would generally accept any method that assesses risks reliably, like having no credit record. Everyone will agree probably, that we are talking about illegal discrimination based on inacceptable grounds, like gender, heritage or sexual preferences. Discrimination based on former ability to pay back loans should be acceptable, also discrimination based on income or presumed job stability, there will be a lot of factors we might accept. If there were factors which looked OK at first, but would in fact be discriminating in favour of straight middle aged white males, we should object to these factors being used, both morally as well as legally. Even with these proxy discriminations ruled out, there will be differences in your groups total loan acceptance rate, but you, as an individual, should get the same evaluation as anyone from any different group in a similar situation.
There is a catch though. What if I told you that formerly, in the good old times, when actual people decided on your loan, these people might have been far more racist, mysogynist, homophobe and otherwise preferential than any current algorithm? I deem it rather probable, that any big company trying to earn money by lending money to you is rather more interested in this said money than pushing some ideology into the world.
On the other hand there is one additional problem, are we able to evaluate the algorithms? Are we able to spot wilfull incorporation of proxy discrimination? Neural nets might be able to produce good evaluation results, usually they are built with no explanatory component. Expert systems are able to explain, but are we able to understand the explanation. Apparently we need better software, designed to be able to prove it is non-discriminatory. Furthermore, we probably should require this software to be transparent, enabling us, as society, if we think the grounds are acceptable. Thus we could lead discussions like this, is it acceptable policy to deny heavy drinkers a loan.
Please have a look at your comment. It appears to be just the communication style which is discussed here, it appears to intend to intimidate people to stop what they are doing. This is not meant funny, I just think we all have adopted this kind of talking way more than we even know.
Do they even know, what they are asking for? If they are absolutely insisting on enforcing the ludicrous "right to be forgotten" a Great Firewall of Europe it needs to be then, as only that will enable the invention of history, which at its core it is.
Fine with me, as long as every contender mentioned in schools also needs to be at least a theory, which would rule out any believe based systems, as these lack two basic prerequisites of scientific theory, verifyability and falsifyability.
There are many people who will value a relationship highly even after it ends, everything you say seems to be negative.
I used a normal generalization where a person would want to keep naked pictures of their ex. If the relationship remains friends after a breakup, it's an exception not the normal. In which case, part of the friendship would entail deleting old naked photo's of the ex out of respect don't you think?
No. Deleting those pictures would be part of trying to get rid of that person. It would be like burning love letters. If you respect a person you should never do that. In addition, I don't think it is a normal generalization but your personal view.
It would be a terrible world indeed where we would stop remembering the good times we had together with some loved one.
If you are a person that has to sit and look at naked photos of the ex to "remember the good times" you either had a relationship that was not really a relationship but a string of sexual encounters, or you are mentally ill. I don't mean that as accusatory, but if the shoe fits...
What makes you denigrate sexual encounters in this way? Speaking only of my own experience I can say that sexuality has always been quite a beautiful part of my contacts with other people. Besides, I did not say that I need those pictures to remember, but I would concur that pictures help to remember, perhaps that is one of the reasons people take pictures in the first place. Furthermore, it is no insult to be mentally ill, try harder.:)
This is probably the core of this court decision: the judges share your opinion. They think taking pornographic pictures is indecent and no modest person should indulge in this. If you look closely you will notice, that this thinking needs some condemation of sexuality itself.
It is possible to think differently, though. I would even go as far as saying: Not taking pornographic pictures is an indicator for a lack of love of the other person or ones own body. As you said, you need to be responsible and talk about how these pictures may be used, but beyond that I cannot see any consequences you should fear.
Have you considered that maybe your view of relationships might not be universal? There are many people who will value a relationship highly even after it ends, everything you say seems to be negative. If a person dwells on a mental image of an ex it apparently has been a time well spent with this person. It would be a terrible world indeed where we would stop remembering the good times we had together with some loved one.
That's a truism, as any believe in anything is based on trust in our own mind, our senses, the trustworthyness of some sources and so on. On the other hand it is definitely not necessary to trust a single entity as the DNSSEC infrastructure, giving each other trust in a hippy fashion by PGP-like infrastructures is always a possibility, even if you are clearly opposed to that.
Any system depending on the integrity of a single entitiy will be measured on the amount of power, social skills or money needed to break this integrity. Is this spreading useless fear? I don't think so.
At least in Germany you have a quite complicated situation, as any contract via fax is valid in principle but will not automatically stand up in court as proof, as it is open to free evaluation by the judge (freie richterliche Würdigung). So, depending on the circumstances, you might not succeed in convincing the judge. Thus, not one of my customers depends on the legal validity of fax, and never has. As email obseleted fax most switched to software based fax or providers offering you a free fax-to-mail gateway. This setup is even less valid as a proof, but it is simply not important. Especially, as nearly noone who is doing business is without email.
In the Thames Valley Police Facebook page you will find a video asking you to inform the police via twitter if you notice anything suspicious.
This does certainly not mean governments would not want the ability to shut down communications, they still might come to the conclusion that shutting down the net or telephones would be a better idea. But things will get interesting the moment politicians actually start to understand the net.
The solution is simple as well, just look at Google and China. Or, more direct, if I want to see the net without German censorship I certainly use google.com, not google.de. If the people of Germany want to use the features, which might no longer be supported on facebook.de (currently under Irish legislation) they will simply switch to facebook.com.
Regarding your problem with identity I must say that I am a strong proponent of clear name policy, but even without Facebook or any other social network will eventually have identity, simply by usage. Trust in governmental or corporate identity control is probably not what we should rely on. With Facebook you and your community will be in control of this information and even disinformation, with restricted databases it will only be the corporations and governments.
In Germany you are allowed to take pictures of any person, you just need their permission to publish or exhibit KunstUrhG 22 with uploading usually being publication. Since 2004 you are not allowed to take pictures of people without their consent inside someones home or similar secured spaces StGB 201a. Outside these special locations you won't need a permission if people are just accessories ("Beiwerk") or part of an assembly ("Versammlung") or parade ("Aufzug") or similar KunstUrhG 23.
Very nice analysis of the german situation and the reasons for the current state of the german privacy discussion.
Nonetheless I think the Facebook bashing in general and in particular the facial recognition feature is moot. As always, the net will route against any obstacles it encounters, no oppressive state can suppress the freedom of information for long. Should germany implement blocking of unwanted content people will start using proxies to access Facebook. If Facebook stops using facial recognition other sites will pop up who will enable people to use it.
An oppressive state will forbid the use of data by the people while using data extensively itself. The question is not if data like (face,name) pairs is collected as anything which is technically feasible and economically viable will be done, the question is Will I, too, be allowed to use this data? I use Facebook, I pay my usage by handing over data on myself, this is a fair trade. In return I am enabled to access data of the people of the world, this will ultimately lead to the disappearance of the misconception of privacy and will be our best bet on fighting oppression.
Our speech is a culturally evolved natural capability of humans. Using speech implies the expectation of an intelligent human as a receiver, alleviating this problem by not trying to create the illusion of intelligence will nevertheless force people to use an intelligent, human ability in a degraded, humiliating way. Not even Artificial Intelligence which will probably create intelligent systems in maybe 30 years will essentially change the situation as it won't be expected behaviour these AI systems will be producing, there is currently no intention of creating anything having all limitations humans have. If, on the other hand, this changes, it might actually work, but you have to expect these systems to really not understand you and be unwilling to help or producing lots of errors if they have a bad day, perhaps because their lover broke up with them or the wages are to low.
Before I was given the proof I could immediately understand ( http://science.slashdot.org/comments.pl?sid=1701394&cid=32729242) I started a little OpenOffice database to give me some experimental evidence. This OpenOffice Base file unfortunately has one error which keeps me from checking bigger datasets but at least it did show me that 100,000 families is not enough. The probability for a brother ranges between 0.42 and 0.51 in my runs.
Slashdot Mirror
There already exist flexible materials which can be made to change shape under an applied current, it should be possible to make them small enough to display haptically at the tip of a gloves finger a fingerprint taken by a fingerprint sensor on the inside of the glove.This would solve the problem of the stolen glove as well as the mistaken belief in biometric access control.
You are perfectly right, with people there is no moral decision involved, they just act, most often inadequately in retrospect, and we are fine with that, as strange as it might seem, with rare cases being "tragic", inescapable, impossible to prevent. This is the problem with algorithms, the accident does not simply happen, it is happening the way it is happening in some explainable way, so people want to know the reasons why it is happening like this and not like something completely different. So it does ot suffice to be better than a human statistically. Don't get me wrong, I firmly believe we can find acceptable moral algorithms and we should pursue creating them.
This set of rules is one ruleset implementing moral judgement. What needs to be done is finding rulesets people could agree upon, which implement acceptable moral decisions and are complete enough to cover even rare situations in a way people are comfotable with.
You are quite right to require non-discrimination, thanks for providing this insight, as I would generally accept any method that assesses risks reliably, like having no credit record. Everyone will agree probably, that we are talking about illegal discrimination based on inacceptable grounds, like gender, heritage or sexual preferences. Discrimination based on former ability to pay back loans should be acceptable, also discrimination based on income or presumed job stability, there will be a lot of factors we might accept. If there were factors which looked OK at first, but would in fact be discriminating in favour of straight middle aged white males, we should object to these factors being used, both morally as well as legally. Even with these proxy discriminations ruled out, there will be differences in your groups total loan acceptance rate, but you, as an individual, should get the same evaluation as anyone from any different group in a similar situation. There is a catch though. What if I told you that formerly, in the good old times, when actual people decided on your loan, these people might have been far more racist, mysogynist, homophobe and otherwise preferential than any current algorithm? I deem it rather probable, that any big company trying to earn money by lending money to you is rather more interested in this said money than pushing some ideology into the world. On the other hand there is one additional problem, are we able to evaluate the algorithms? Are we able to spot wilfull incorporation of proxy discrimination? Neural nets might be able to produce good evaluation results, usually they are built with no explanatory component. Expert systems are able to explain, but are we able to understand the explanation. Apparently we need better software, designed to be able to prove it is non-discriminatory. Furthermore, we probably should require this software to be transparent, enabling us, as society, if we think the grounds are acceptable. Thus we could lead discussions like this, is it acceptable policy to deny heavy drinkers a loan.
Please have a look at your comment. It appears to be just the communication style which is discussed here, it appears to intend to intimidate people to stop what they are doing. This is not meant funny, I just think we all have adopted this kind of talking way more than we even know.
Do they even know, what they are asking for? If they are absolutely insisting on enforcing the ludicrous "right to be forgotten" a Great Firewall of Europe it needs to be then, as only that will enable the invention of history, which at its core it is.
Fine with me, as long as every contender mentioned in schools also needs to be at least a theory, which would rule out any believe based systems, as these lack two basic prerequisites of scientific theory, verifyability and falsifyability.
Could it be that this is indicating society has to evolve, allowing more freedom of personal expression rather than trying to keep the old ways?
There are many people who will value a relationship highly even after it ends, everything you say seems to be negative.
I used a normal generalization where a person would want to keep naked pictures of their ex. If the relationship remains friends after a breakup, it's an exception not the normal. In which case, part of the friendship would entail deleting old naked photo's of the ex out of respect don't you think?
No. Deleting those pictures would be part of trying to get rid of that person. It would be like burning love letters. If you respect a person you should never do that. In addition, I don't think it is a normal generalization but your personal view.
It would be a terrible world indeed where we would stop remembering the good times we had together with some loved one.
If you are a person that has to sit and look at naked photos of the ex to "remember the good times" you either had a relationship that was not really a relationship but a string of sexual encounters, or you are mentally ill. I don't mean that as accusatory, but if the shoe fits...
What makes you denigrate sexual encounters in this way? Speaking only of my own experience I can say that sexuality has always been quite a beautiful part of my contacts with other people. Besides, I did not say that I need those pictures to remember, but I would concur that pictures help to remember, perhaps that is one of the reasons people take pictures in the first place. Furthermore, it is no insult to be mentally ill, try harder. :)
This is probably the core of this court decision: the judges share your opinion. They think taking pornographic pictures is indecent and no modest person should indulge in this. If you look closely you will notice, that this thinking needs some condemation of sexuality itself.
It is possible to think differently, though. I would even go as far as saying: Not taking pornographic pictures is an indicator for a lack of love of the other person or ones own body. As you said, you need to be responsible and talk about how these pictures may be used, but beyond that I cannot see any consequences you should fear.
Have you considered that maybe your view of relationships might not be universal? There are many people who will value a relationship highly even after it ends, everything you say seems to be negative. If a person dwells on a mental image of an ex it apparently has been a time well spent with this person. It would be a terrible world indeed where we would stop remembering the good times we had together with some loved one.
Sorry but you got to trust someone.
That's a truism, as any believe in anything is based on trust in our own mind, our senses, the trustworthyness of some sources and so on. On the other hand it is definitely not necessary to trust a single entity as the DNSSEC infrastructure, giving each other trust in a hippy fashion by PGP-like infrastructures is always a possibility, even if you are clearly opposed to that.
Any system depending on the integrity of a single entitiy will be measured on the amount of power, social skills or money needed to break this integrity. Is this spreading useless fear? I don't think so.
At least in Germany you have a quite complicated situation, as any contract via fax is valid in principle but will not automatically stand up in court as proof, as it is open to free evaluation by the judge (freie richterliche Würdigung). So, depending on the circumstances, you might not succeed in convincing the judge. Thus, not one of my customers depends on the legal validity of fax, and never has. As email obseleted fax most switched to software based fax or providers offering you a free fax-to-mail gateway. This setup is even less valid as a proof, but it is simply not important. Especially, as nearly noone who is doing business is without email.
Done in 2007: https://extras.skype.com/486/view
In the Thames Valley Police Facebook page you will find a video asking you to inform the police via twitter if you notice anything suspicious.
This does certainly not mean governments would not want the ability to shut down communications, they still might come to the conclusion that shutting down the net or telephones would be a better idea. But things will get interesting the moment politicians actually start to understand the net.
The solution is simple as well, just look at Google and China. Or, more direct, if I want to see the net without German censorship I certainly use google.com, not google.de. If the people of Germany want to use the features, which might no longer be supported on facebook.de (currently under Irish legislation) they will simply switch to facebook.com.
Regarding your problem with identity I must say that I am a strong proponent of clear name policy, but even without Facebook or any other social network will eventually have identity, simply by usage. Trust in governmental or corporate identity control is probably not what we should rely on. With Facebook you and your community will be in control of this information and even disinformation, with restricted databases it will only be the corporations and governments.
In Germany you are allowed to take pictures of any person, you just need their permission to publish or exhibit KunstUrhG 22 with uploading usually being publication. Since 2004 you are not allowed to take pictures of people without their consent inside someones home or similar secured spaces StGB 201a. Outside these special locations you won't need a permission if people are just accessories ("Beiwerk") or part of an assembly ("Versammlung") or parade ("Aufzug") or similar KunstUrhG 23.
Very nice analysis of the german situation and the reasons for the current state of the german privacy discussion.
Nonetheless I think the Facebook bashing in general and in particular the facial recognition feature is moot. As always, the net will route against any obstacles it encounters, no oppressive state can suppress the freedom of information for long. Should germany implement blocking of unwanted content people will start using proxies to access Facebook. If Facebook stops using facial recognition other sites will pop up who will enable people to use it.
An oppressive state will forbid the use of data by the people while using data extensively itself. The question is not if data like (face,name) pairs is collected as anything which is technically feasible and economically viable will be done, the question is Will I, too, be allowed to use this data? I use Facebook, I pay my usage by handing over data on myself, this is a fair trade. In return I am enabled to access data of the people of the world, this will ultimately lead to the disappearance of the misconception of privacy and will be our best bet on fighting oppression.
Our speech is a culturally evolved natural capability of humans. Using speech implies the expectation of an intelligent human as a receiver, alleviating this problem by not trying to create the illusion of intelligence will nevertheless force people to use an intelligent, human ability in a degraded, humiliating way. Not even Artificial Intelligence which will probably create intelligent systems in maybe 30 years will essentially change the situation as it won't be expected behaviour these AI systems will be producing, there is currently no intention of creating anything having all limitations humans have. If, on the other hand, this changes, it might actually work, but you have to expect these systems to really not understand you and be unwilling to help or producing lots of errors if they have a bad day, perhaps because their lover broke up with them or the wages are to low.
Before I was given the proof I could immediately understand ( http://science.slashdot.org/comments.pl?sid=1701394&cid=32729242) I started a little OpenOffice database to give me some experimental evidence. This OpenOffice Base file unfortunately has one error which keeps me from checking bigger datasets but at least it did show me that 100,000 families is not enough. The probability for a brother ranges between 0.42 and 0.51 in my runs.