Breakout II.a (Bren 1424): "Information Credibility"
- Facilitators
- Participants
Issues of information quality, reliability, authority, reputation, provenance, and accountability are increasing important in the current era of social computing. In the age of Web 2.0, traditional notions of credibility premised on hierarchically-organized and centralized authority become problematic. Public knowledge-production websites do an end-run around traditional accreditation methods by evolving elastic combinations of programmatic, semi-formalized, and informal distributed-authority mechanisms that call for a new understanding of credibility and authority. (See the work of the Credibility and Digital Media Project at UCSB, co-directed by Transliteracies faculty Miriam Metzger and Andrew Flanagin.)
- What are the top-ten research problems/opportunities in this area?
- And is there a recommendation for a particular issue for RoSe to pursue?
Comments (2)
miriam metzger said
at 4:55 pm on Feb 26, 2010
To what extent are authority and credibility related? Is deriving the chain of authority (source) for some information a precondition of credibility?
RoSE is a melding of wikipedia and a social network, each of which have different models for determining credibility (credibility in databases is different from credibility in social networks). Which elements of both forms of credibility evaluation would be most helpful in looking at issues of credibility in RoSE?
What metadata would we want to help users of RoSE evaluate the credibility of information within it? (1) information about texts, (2) information about authors, (3) comment page users' opinions about credibility/ratings (but expertise is topically bounded)
Who contributes info? Anyone, experts only, something in between (e.g., anyone can contribute but all entries vetted by topical experts)? Pros and cons?
No credibility system will satisfy everyone's needs?
To be continued...I have to leave now :(
miriam metzger said
at 3:57 pm on Mar 5, 2010
Continued...
To what extent can information provenance (i.e., making the source of information traceable) solve the problem of credibility in collectively-authored and edited environments?
Where does cultural authority fit into RoSE? Is RoSE offering another model of cultural authority, or can RoSE build in cultural authority?
Will humanities research be transformed or enhanced by RoSE?
Distinct communities have their own means and norms for evaluating credibility, how can we design this into a system like RoSE? Should we design it in, or should methods of evaluating credibility in RoSE just evolve on their own?
We discussed the very difficult problem that credibility is subjective. What one person believes as credible may not be taken as credible by another person. This lead to the caution that any credibility scheme would need to be flexible.
We talked about building credibility systems that are robust to collusion, using tags, developing authentication and vetting, restricting information contributors to "experts" only, implementing some sort of rating system with a feedback or "rate-the-raters" feature built in and/or weighting ratings by expertise. We also discussed the downsides to all of these ideas.
You don't have permission to comment on this page.