Panagiotis Takis Metaxas (at the Berkman Center) and Eni Mustafaraj have written a paper called “trails of Trustworthiness in Real-Time Streams” [pdf] about how to support critical thinking about social networking conversations, while maintaining privacy. From the abstract:

When confronted with information that requires fast ac- tion, our system will enable its educated users to evaluate its provenance, its credibility and the independence of the multi- ple sources that may provide this information.

They say the only real hope is to solve the problem within closed streams that provide membership functions because there “it is possible to determine the a priori trustworthiness of a message received,” by evaluating the credibility of users on particular topics. They believe this can be done by watching the actions of users. For example, “In general, the more often a user re-posts messages from a sender, the more trusted the sender becomes.” And: “A message that has been sent by different, independent users has more trustworthiness than one that has been initiated by a single user.”

There’s much more in their paper…