Pre-class Discussion for Jan 17
Jump to navigation Jump to search
- If your copy of Wisdom of Crowds, Chapter 1, has an illegible page 10 or page 11, try the Amazon Online Reader for the hardcover edition. I searched for a phrase on page 10 and read just those two pages online. Jumpingdeeps 21:31, 16 January 2008 (EST)
Jonathan Zittrain, Future of the Internet, Chapter 9, pp. 221-234
- I thought ReputationDefender services was a good idea, so looked online to find some more information. It was interesting to see one of AutoAdmit's responses to the ReputationDefenders. I am not sure how much of what Mr. Cohen asserts is true, however, it addresses cyber-harrassment and makes a proposal to ReputationDefender. KStanfield 22:29, 16 January 2008 (EST)
- Perhaps an additional method of saving some shred of the privacy idea, could be to create a new form of privilege, or rule of evidence, to prohibit introducing evidence of online activity obtained from tracking by ISPs or search engines. While this obviously wouldn't do much to truly protect privacy, in the sense that others, including the government and law enforcement, could still obtain this information. But, it would at least some semblance of privacy that at least internet activity couldn't be used against you in court. This might at least lessen the sort of Big Brother concerns. The argument against a kind of privilege, however, is that other forms of privilege are designed to encourage some kind of other desirable activity, such as talking to your attorney, whereas this privilege seems that it would only encourage illegal activity on the internet. --Kgrose 00:13, 17 January 2008 (EST)
- The privilege idea, while lessening Big Brother concerns, does nothing in the private sector. Privacy can still be invaded by online tracking or through various other means. One's reputation could still be destroyed by an unfortunate moment in public. I agree with the idea that this will likely result in such moments becoming of considerably less importance in the future as society gets used to this and it becomes even more widespread. The problem, however, is that doesn't protect those of us that are growing up in this era, who have to deal with potential employers that aren't used to the technology yet, but are exposed to it. I was specifically reminded of the warning a group of us received from, I believe, Dean Cosgrove, at a Bar Exam info session last semester (as many of you probably received as well). Basically the idea was that we need to be wary of what we have on our Facebook and MySpace pages because the state bar will sometimes view them in evaluating whether you have good moral character. Needless to say, I think many found this revelation somewhat disturbing. State bars could certainly also do a Google search or look at Youtube clips if an applicant's name was attached. It seems incredibly unfair to perhaps prevent an individual from entering their career due to some prior indiscretion. While in the future, state bars may understand this problem and ignore isolated incidents, it certainly doesn't solve the problem for current students. Do others agree that this is a serious problem? If so, any ideas for solutions? --Kgrose 00:13, 17 January 2008 (EST)
- Is reputation bankruptcy likely to be effective? I agree with JZ that if we move to a world where reputation is completely important, a method of forgiveness or penance would be valuable. But since, as JZ notes, information is hard to get rid of once it is on the Web, I find it hard to believe that any network could impose a working forgiveness mechanism. Just as WikiScanner created a big to-do, I can imagine sites that reveal reputation elements people have deleted to be very popular. A third-party site that cached 'reputation' and revealed when the cached reputation was different from the on-site reputation could easily undermine the bankruptcy system, while attracting visitors by revealing information that 'information debtors' themselves would reveal as especially sensitive. Dankahn 01:25, 17 January 2008 (EST)
- My impression of the solutions suggested in this chapter was that they might help *me* regulate the information that *I* put on the web. For example, I could use robot.txt or metadata tagging of pictures that I upload to limit the extent to which my information can be indexed or copied. I fail to see how this helps in situations where X puts something on the web about Y, which seems to be the true concern in Privacy 2.0. Am I missing something? For example, if Jane Doe took a picture of John Smith at a rally, Jane would (presumably) control robot.txt settings and metadata tagging. Would John Smith be able to use reputation bankruptcy or some other means to disassociate himself from the picture? If not, the Privacy 2.0 problem still exists. If so, what about Jane Doe's property claim over the picture she took? Cjohnson 12:26, 17 January 2008 (EST)
Lior Strahilevitz, 'How's My Driving?' for Everyone (and Everything?)
- As noted in my previous post when Prof. Zittrain discussed "How's My Driving," I am skeptical of using this reputation system for all non-commercial vehicles. I believe that it is a great program in the sense that it may reduce accidents, decrease aggressive driving, and create more courteous drivers. However, I am still not sold on the idea. Although we were not given the article in its entirety, I still have problems with abusive behavior, over-reporting of negative behavior (a concern mentioned by another student in the previous post), and multiple point deductions for one act. Abusive behavior and over-reporting were mentioned in the previous posts, so I won't mention it here. I am least concerned about the multiple reporting for the same act, since I believe there are possible methods of correction. Another point in "How's My Driving" that concerned me was the reporting for non-illegal behavior. Prof. Strahilevitz specifically states that "...illegal driving behavior annoys research subjects much less than various hostile gestures and discourtesy on roadway" (FN 123). Should someone receive point deductions for not being a nice person? Although I do not condone discourteous behavior and think that eliminating aggressive driving is a positive, why should we force people to pay fines for being discourteous if they did not break the law? It's possible that such behavior would not lead to point deductions and therefore would not effect the drivers, but then what would discourage this behavior? I do believe this system could ultimately work if applied correctly. However, I would like to see more studies before even playing with the idea of its implementation. Reputation systems may work on e-Bay and in other contexts. Will it really work with non-commercial vehicles? Even if there are positive results for commercial vehicles, installation of technological devices will make it much easier for people to report. Thus, the possibility of abuse is much greater. KStanfield 22:15, 16 January 2008 (EST)
- Although I think this type of reputation system does work well online in certain instances (like eBay), I am very concerned with the potential effects of a "How's My Driving?" system being implemented. Logistical difficulties aside, I think the more fundamental question is whether we as a society really want to start a precedent in which people report on each other? Even if itâs not required and only certain people do, doesnât this create a cultural norm that we should be hesitant of? Anna 22:24, 16 January 2008 (EST)
- I agree with Anna that there are some serious cultural norm issues. How often do people tell their kids not to be a tattle tale? And now we're thinking about implementing it in widespread fashion to solve bad behavior on the roads and on the internet? One particular way, however, in which I thought online ratings could make sense and do a lot to stop the so-called bad bits, is a rating attached to any IP address from which you download a file. This way, before you download a file that could be a potential virus, you can check that person's rating and even know if they have ever sent a virus before. I liked the informational asymmetry argument posed by Strahilevitz and it would apply here as well. In downloading bad files, the problem is often that we don't know the sender. A rating system would correct that to some extent. --Kgrose 00:33, 17 January 2008 (EST)
- I also agree with Anna and Kgrose that we should be more concerned with the social norms we are creating if we impose "How's My Driving." I feel like the role of social norms might sometimes not get enough credit in this area. For instance, in the Kathy Sierra, it seems is possible that one of the (many) problems is that there is a general sense that harms online are not as bad as harms 'in real life.' Beyond just feeling safer due to the anonymity of the Internet, the attackers might have thought what they were doing was not as bad becasue 'it was just online.' If we could cultivate a stronger social norm that harms online can be just as bad as harms in 'real life,' perhaps fewer Sierra-type incidents would occur. While there creating a social norm could improve the situation, here perhaps, as Anna and Kgrose point out, we should be wary of creating a social norm with harmful effects. Dankahn 10:58, 17 January 2008 (EST)
- Anna makes a good point about the social norms this creates. Besides the social norms, though, I feel that there is a huge Big Brother element to the system that Strahilevitz proposes. Essentially, your every minute on the road is going to be watched by multiple actors from multiple angles, and you become vulnerable to being reported to authorities that have the power to significantly affect your driving activity. In this sense, I wonder if the eBay feedback process is really an apposite example. Even with eBay, I've heard of vendors that will offer a disgruntled user lots of freebies to prevent him/her from giving the vendor a bad rating, and of users who exploit these weaknesses. In addition to making all of us reporters/tattletales, the HMDFE system would put us in the vulnerable vendor position, with greater real-life consequences than a mere percentage drop in one's approval rating. --K.lee 12:08, 17 January 2008 (EST)
- I think that the social norms are less of a concern in an area like driving that is already very public. We're already surrounded by other drivers on the roads; the only difference that HMDFE would make is giving those other drivers an opportunity to effectively comment on our actions. The analogy to the citizen-informers of the Stasi is clear, but in this area at least, relatively innocuous.
- I'm also concerned about the implementation of the system in terms of multiple points for certain infractions or a bias towards negative feedback or against certain marginalized groups, but I think this is more a problem of the method of implementing the system than a problem with the system itself. Certain fixes to the system--for example, setting point values for certain actions, or devising a system that would correlate user feedback with the characteristics of the rated drivers--might help to mitigate these problems. Eroggenkamp 14:07, 17 January 2008 (EST)
- Another problem with this system is that reviewers may rate other drivers poorly based on their own prejudices towards them. The cumulative effect, then, may be that some traditionally marginalized groups will receive poor ratings despite driving no differently from anyone else. In order to correct for this, the system may have to pay attention not just to the reviews the drivers received, but also the ratings that drivers give out. It would add significant complexity to the system to monitor for consistent prejudice in the ratings that drivers gave. To this extent, then, a system of speed cameras like those adopted in England may be superior. While those clearly have many privacy concerns associated with them, they do seem to save on enforcement costs about as much as an HMD system would, and they have the added advantage of being unbiased. --NikaE 11:21, 17 January 2008 (EST)
- I agree that there are potential abuses when we give drivers the discretion to review other drivers. Especially problematic in my opinion was Strahilevitz's suggestion of fining bad drivers and monetarily rewarding good drivers according to the feedback received on their driving, especially if such a system was implemented in lieu of or to supplement professional policing. I am worried that such as system would create a perverse incentive for wealthy drivers to ignore traffic rules because they can easily pay fines (although I have to admit the consequences for minor traffic violations are fines as well, but greater trangressions sometimes include traffic school, and repeat offenders will have their license suspended - which was not an option discussed in this reading). I also feel that giving people the discretion to choose the amount of points to deduct from a fellow "bad" driver would also be prone to abuse by someone who is having a bad day, or holds certain prejudices against whatever characteristic the driver may posses (as many other people have commented on today and in previous discussions). Nevertheless, I think these concerns are specific to Strahilevitz's suggestions and HMDFE. I think anonymous peer reviewing is extremely useful in the eBay context because (1) it doesn't include money, and (2) the users do not see each other and a relatively anonymous so their judgments may be less based on social prejudices and more so on actual interactions with the other user. And with something as large and heavily populated as the internet, it would be impossible to restrict policing to official bodies and still be effective. User participation and reviewing may indeed be a suitable cost-effective compliment. Cseif 13:50, 17 January 2008 (EST)
- I am also a little doubtful about the argument that HMDFE would reduce police monitoring costs, as the system seems to spawn a number of its own problems. For instance, while we weren't assigned the section about ameliorating the problem of false and malicious feedback, one can foresee significant costs associated with verifying driver reports (the fact that one's own vehicle would be recognized in the reporting process doesn't seem like enough of a deterrent to false reports) or preventing tit-for-tat reports where angry drivers keep calling to tell on each other. Also, part of the reason HMD call centers have such low call volume may be that people are not really cognizant of the effects of their calls and are just doing it for the expressive benefits. Once implemented on the scale that the article proposes, combined with the very real effects of civil fines/penalties, call volume would increase dramatically and governments would have to devote more resources to maintaining the centers. --K.lee 12:32, 17 January 2008 (EST)
- I agree that HMDFE is problematic in many ways. But to look at things from the other side...
- Perhaps the system would benefit from the "wisdom of the crowds." Malicious reporting that does not get filtered out by built-in mechanisms could amount to nothing, when the collective average is what determines the end result.
- The author does not seem to suggest huge burdens on bad drivers - rather, small civil penalties (bad behavior) and rewards (good behavior.) Does this mitigate concerns about HMDFE having a significant impact on people's driving activities?
- For simple +1/-1 reporting, automated call centers could keep the costs of the system low.
- While reviewers might be influenced by personal prejudices, so are police in our current system. Perhaps democratization of the process would mitigate, rather than exacerbate, this problem.
- Perhaps HMDFE would suffer from *under*-reporting. If everyone is driving 75 when the speed limit is 65, can they really be expected to report each other? Or report someone going slightly faster at 80, so long as that person is not behaving recklessly? Cjohnson 12:50, 17 January 2008 (EST)
- Assuming the correlation between accidents and reckless behavior is stronger than that between accidents and speeding alone then that would be a good thing, especially compared to the current regime of overenforcing speeding merely because it's conspicuous. Jason 13:59, 17 January 2008 (EST)
James Suroweicki, Wisdom of Crowds, Chapter 1
- Does Suroweicki ignore the dark side of crowds? In particular, in the Kathy Sierra incident, much of the blame went to 'mob' mentality. What is the difference between a 'mob' and a wise crowd? Is it based on context? Is the difference that the wise crowds Surowiecki describes work in isolation, with their efforts aggregated, while 'mobs' act together? If the last suggestion is accurate, does it suggest that efforts to govern through deliberative processes are misguided? Dankahn 01:07, 17 January 2008 (EST)
- The Federalist Papers explain how the U.S. government is structured to restrain the majority will from being imposed too quickly. If such against popular 'passions' are valuable in the area of government structure, in what contexts are such checks and balances wise, and in which do they inhibit relying on the wisdom of crowds? Dankahn 01:07, 17 January 2008 (EST)
- Does Suroweicki ignore the emerging literature on systematic biases in human thought? Psychologists, economists, and biologists such as Kahneman and Tversky have developed with increasing sophistication insight into common cognitive biases that pervade human thought.  In situations in which cognitive bias is likely, isn't relying on crowds unwise? Would it not be better to rely on individual experts, trained to avoid common biases? Dankahn 01:07, 17 January 2008 (EST)
- I do think Suroweicki did a poor job of emphasizing the factors that make crowds more likely to come to good decisions than bad decisions. The "conditions that characterize wise crowds" are tossed off in a paragraph on page 10 that gets very little further discussion (at least in our excerpt). Suroweicki notes that diversity of opinion, independence, decentralization, and aggregation are keys to good group decision making, but I think that these characteristics are rarely present in certain forms of decision making. Eroggenkamp 14:01, 17 January 2008 (EST)
- Is the aggregation and averaging of isolated individual (good point above re: crowds vs. mobs) efforts likely to mitigate rather than amplify the effects of such biases? Jason 11:48, 17 January 2008 (EST)
- Note an intersting story about potentially unethical uses of the wisdom of crowds. In 2003, DARPA, an arm of the Pentagon (that happens to have a big connection to the history of the Internet Arpanet on Wikipedia) was found to have a plan to develop a prediction market for terrorism attacks. A public outcry developed, and the plan was cancelled. Does the public outcry, ironically, illustrate the lack of wisdom of crowds? Did political outcry kill what could have been an effective means of predicting and preventing future terrorist attacks? Or does the offensiveness of allowing gamblers to profit off of terrorist attacks outweigh the potential benefit? Dankahn 01:07, 17 January 2008 (EST)