Professional Confidentiality and Privilege in an Investigation
CrowdFlower is a young company that engages armies of Internet workers to perform repetitive tasks, in exchange for small per-task fees or rewards. For example, CrowdFlower workers might one-by-one evaluate tens of thousands of Twitter tweets that mention the brand name of a corporate client.
CrowdFlower believes its services could help assess large quantities of e-documents in an investigation, such as a lawsuit or inquiry by a government regulatory authority. For example, in response to a subpoena from the federal Financial Crisis Inquiry Commission (FCIC), investment banker Goldman Sachs unloaded a billion (!) pages of legal records on the government. CrowdFlower offered its services. It offered to deploy an army of virtual workers to sort the records.
That offer caught my attention, because I see that the quantities of records in modern investigations are skyrocketing. I agree with CrowdFlower that crowdsourcing could become an important tool of e-discovery and electronic records management.
Therefore, I inquired whether CrowdFlower could demonstrate its technology at the SANS Institute’s September 27-28, 2010, Las Vegas Summit on E-Records Policy. The company responded favorably, and it has teamed with an eDisovery vendor to assist with the demonstration. Then it asked me how to assemble a presentation.
The answer is that I don’t entirely know yet. So I’m blogging here, to solicit input from you, my reader.
My guess is that CrowdFlower could do this: It could gain access to a large group of sample records, such as government records from someone like the Sunlight Foundation. Then, it could devise a survey of the records for its worker army to complete. Before CrowdFlower presents at the Summit, it could begin the survey and compile/analyze results. Then, in its Summit presentation, CrowdFlower could explain the results, and even demonstrate . . . live . . . how its worker army is continuing to survey the records.
CrowdFlower is not the only company engaging legions of Internet workers. CrowdFlower’s distinguishing secret sauce is its quality-control technology (algorithms and procedures) for filtering and evaluating worker activities so that one can have confidence in the overall results of a work project.
As a lawyer who cogitates on the future of e-discovery, I can imagine the use of crowds to sift through massive volumes of records in lawsuits. However, such records often need to be kept confidential. To achieve confidentiality (and perhaps even the protected status of attorney-client privilege and attorney work product), I envision someone like CrowdFlower enlisting large numbers of professionals who have an ethical duty of confidentiality and/or a reputation for confidentiality. Such professionals might be (a) licensed (credentialed) professionals such as lawyers, paralegals or certified public accountants, or (b) other people who have a track record for trustworthiness.
Dear Reader: What do you suggest that CrowdFlower do at the SANS E-Records Summit (Las Vegas)?
Crowdflower's demonstration will be a highlight of the Summit.
--Benjamin Wright, Senior Instructor, SANS Institute
Lawyer-Swarm I have an update on the crowdsourcing ideas I expressed above: In the Lehman Brothers bankruptcy, investigators have reviewed astronomical numbers of electronic records such as email and instant messages -- the equivalent of 34 million pages of paper! How did they do that? They swarmed the records. The law firm leading the effort, Jenner & Block, hired "teams of contract lawyers to comb through the pages." "Lehman Brothers: how investigators sift through e-mail," Financial Times, July 22, 2010, p. 10 (US edition).