CLIP Lab: Doug Oard (iSchool/UMIACS)

Title: Information Abolition

Abstract: In the field of information retrieval, we take it as our goal to help people find what they want to see.  In this talk I will argue that it is high time that we also begin to formalize our thinking about what I will call “information abolition,” by which I mean preventing people from finding things that they should not see.  Back when information was scarce, we could do this just by keeping things that should not be seen out of our index.  But today it is increasingly common to find important information (i.e., information that should be found) intermixed with sensitive information that needs to be protected for one reason or another (personal privacy, commercial interests, national security, ...). As with every problem that we work on, this is not a matter of all or nothing.  Rather, the challenge is to balance the risks.  Our first goal, therefore, must be to articulate what it would mean to succeed at an information abolition task.  I’ll begin the talk by reviewing a few information abolition tasks, followed by a deep dive into some recent research on one such task – the protection of privileged content in civil litigation.  I’ll then draw on that story to begin to lay out a structure for thinking about evaluation of information abolition, and finally I’ll conclude by articulating a few challenges that I think we will need to address as a research community as we tackle an increasing range of information abolition problems.