Security weblog
 Friday, May 17, 2002
Digital identity and privacy
The major constituencies involved in a privacy-enabling protocol or system must be willing to sacrifice the information that could be collected about the other parties or their inputs. In the absence of legal requirements. that are generally understood, technologically feasible, and consistently enforced, use of such protocols and systems must be voluntary and bilateral. However, in e-commerce transactions, these constituencies have conflicting interests and asymmetric power. Why should a powerful content/service provider wantingto learn information about his users agree to run a protocol that deprives him of this very information? Industry is likely to the follow the "Know your customer" mantra.
Assuming that corporate entities make decisions motivated primarily by profit (and that a good reputation for respecting customers' privacy has a measurable positive impact on pro tability), these entities should only switch infrastructural technologies if the expected net present value of the benefits of switching is greater than its costs. Experience shows that this makes infrastructural switching rare, slow, and painful.
Often, part of what makes a business an Internet business" is that it can use pre-existing Internet infrastructure to get a cost advantage over its competitors. If privacy technologies require widespread infrastructure redesign, they vitiate this principle of Internet business success, and content/service providers probably will not adopt them.
Perhaps a reason for the universal identity infrastructure to have privacy supporting cabilities built in so that the infrastructure switch means only small change in config files. 11:53:28 PM 
Privacy, online and offline identities
While market forces might ensure fair use of data connected to the on-line identity of individuals (with common satisfaction for all the parties), they evidently do not guarantee optimal use and appropriate protection of the off-line identity (with detrimental economic consequences). Information technology can be used to split those identities (or make the linkages between the identities of an individual too costly for any practical application), but without economic incentives no technology can reach widespread adoption. The need arises for an additional contribution. In particular, legal intervention, on the model of the EU directive on data protection, or as proposed in Samuelson (2000), could put constraints and liabilities on the side of the parties receiving private information. Such constraints should be calibrated to compensate the moral hazard and asymmetric information in the market of personal data. By generating incentives to handle personal information in a new way, the interventions might allow the growth of the market for third parties providing solutions that anonymize off-line information but make it possible to share on-line profiles. If privacy is a holistic concept, only a holistic approach can provide its adequate protection: economic tools to identify the areas of information to share and those to protect; law to signal the directions the market should thereby take; and technology to make those directions viable.
Different requirements for protection of off-line (or Tier 1) and on-line (Tier 2) identity details stemming from overall utility function is an argument, I have never heard before. 11:25:37 PM
|
|
|
