When we meet someone there is a complex dance whilst we “get to know them”. One of the core activities during this period is generating trust between oneself and the other person. Self-disclosure plays a large part in creating this trust. Self-disclosure is the act of telling people about yourself; disclosing information to the other person. Initially these disclosures are small: musical tastes, favourite films and so on. As we gain more trust in the other person, these disclosures become more personal; say, a problem at work.
In making these disclosures, we are showing the other individual we are placing trust in them, which reassures them they can place their trust in us. In addition to building trust, self-disclosure is a reciprocating and reinforcing gesture which allows a kind of self-righting system; if someone is unwilling to disclose as much to me as I to them, something must be amiss. You can read a more detailed discussion of this here
The internet is still a fairly anonymous place. Though people are tending towards using their real identities when posting online now, it is still relatively easy to create some fictitious persona and act through that. In some circumstances this is a bit of fun, say in World of Warcraft, but often the simplicity in creating a fake persona is a significant barrier to creating trust between two parties when it is required, such as when making an online transaction through ebay.
Recently, especially with the explosion of social networking sites such as myspace and facebook, the number of people with a page on the internet which describes (approximately!) their real selves, rather than an invented persona, is growing markedly. Often these sites encourage significant self-disclosure—indeed this has lead to problems between employers and employees. Other modes of self-disclosure online include blogs, forums, twitter and many others. Is it possible to use these instances of self-disclosure to allow for a trust relationship to form between two people without their actually communicating with one another?
I’ve been writing this blog for at least five years now; I would believe there is a high-level of self-disclosure when taken in totality. If someone has been reading this site for a number of years, it would be interesting to know whether they would trust me more, and to what extent, than someone they had no exposure to: would they trust me more on ebay than another new seller? (I’d like to think so, but have no empirical evidence for this.)
Even if the guess that the self-disclosures I’ve made over the years does inspire trust is correct, there are still problems.
First and foremost is I have no simple, autonomous way to link one identity with another, say my ebay identity with my website, at least in a reliable way which would prevent another person doing the same thing (as a simple link would not). I do not know whether MicroID is good enough for this purpose, but it is the kind of technology which could be.
The main criteria for solving this problem is the need for a very high probability that the sites you are using to build trust really belong to the party you are seeking to build trust with. I think this technical problem is surmountable.
A second problem is discovering just what sorts of information are most valuable in building trust in different scenarios. The sort of information required for trust in an ebay transaction probably differs from that needed for trusting some program I write will not destroy your computer and this again may differ from that which would be required to trust my advice given in a forum post. I am not sure what the solution to this is.
Following from this is which sorts of interaction are amenable to trust through self-disclosure vs other forms, such as trust by reputation. Of course, different forms of trust building can be complementary. This is more of a societal question than a technological one (as many of the most difficult ones are!). A further related point is how much trust can be built without revealing very sensitive information, which is impossible in the public disclosure under discussion.
Thirdly this methodology provides no guarantee I will follow through on my promises; instead it provides a route to increasing your trust of me, which is a predictor of the probability I will follow through on my promise. In computing we often wish for complete guarantees rather than probabilities. In real life, however, our relationships are based on probabilities rather than certainties1. As this proposal is addressing person-to-person relationships, or person-to-business, I imagine a probability-based system will suffice. Therefore, I don’t believe this would cause a self-disclosure based trust system to fail.
A fourth problem is that of relative power affecting how much trust self-disclosure engenders. If I am dealing with a person of similar power to myself, then the amount I can place trust in someone who self-discloses is greater than if I am dealing with an entity of far greater power than myself. If a police officer discloses his identity to me, I trust him less than I would a person of similar power doing the same thing; the policeman has such greater power over me that he has little to loose from showing me his ID vs someone else.
It could be, then, that under certain circumstances self-disclosure doesn’t work as a policy for building trust. This does not destroy the concept, however, as there are many situations where increased trust between peers would be beneficial on the internet.
People are building up presences related to their real identities on the web. It would be useful if these presences could be linked together to allow for increased trust between people who are not connected in real life, especially for peer-to-peer interactions such as those between a buyer and seller on ebay. Online presences often involve forms of self-disclosure which is used to build trust in real life. Therefore, if we can provide a technological basis for joining presences across sites into a single identity, we allow for more meaningful trust-relationships to be built up, which will provide a major step forward for interactions on the internet.
1 Even when signing contracts, we are assuming a high probability the other person will not break contract due to the punishments in the contract rather than a certainty