IQPC eDiscovery Panel-Protecting Privileged Communication

by Cary J. Calderone, Esquire

Moderator, Mark Michels, Managing Attorney, Cisco Systems, Inc.
Craig Carpenter, Vice President and General Counsel, Recommind, Inc.
Martin T. Tulley, Partner and E-Discovery Practice Chair, Katten, Muchin, Rosenman, LLP

This panel was focused on Federal Rule of Evidence (FRE) 502 which governs attorney-client privilege and work product; limitations on waiver.  502 was drafted specifically to cover electronic discovery and inadvertent disclosures.  I have had the pleasure of listening to Mark Michels at other presentations and, as an audience member, I always appreciate that he tries to make his panel discussions interesting, lively, and a little bit fun.   After sitting in on 5 or 6 eDiscovery sessions over the two days at the IQPC eDiscovery conference, I believe anybody sitting in this audience appreciated that he made the panel discussion entertaining, as well as informative.

The first hot topic was protective orders with claw-back provisions. The panel was interested in whether people had ever had a "quick peak" (i.e., noticed that your opposition sent you privileged information and had to report it to them) and made the following points:

• Mistakes will happen.  We are dealing with an enormous volume of data.  There is fear of privilege waiver but there is also significant cost in taking precautions. 
• FRE 502- Now with 2 years of precedent it is better understood.
• 502(d) limits waivers of attorney/client privilege, promotes certainty and reduces litigation risks.
• 502(e) indicates that a Protective Order is still prudent.
• There is a big presumption against broad subject waiver.

Mark asked, "when is a disclosure not a waiver of privilege?"
Martin provided a checklist of factors:

• Privilege?-Was it privileged to begin with?
• Inadvertent?-Was the disclosure inadvertent?  Some courts not-intentional equals inadvertent. Other courts have a checklist of factors to consider.
• Reasonable?-Advisory committee notes to 502 do not define it.  Rather, the Committee notes list factors to consider such as precautions and steps taken after disclosure to attempt to rectify?
• Extent of the disclosure?-Overriding issues of fairness.  Was there a defensible Records Management program?  For example, was it 4 out of 10,000 documents?  Did the disclosing party take quick action?  Court found it reasonable and therefore there was no waiver of the privilege.  
• Best practice-Always have a protective order with a claw-back provision.
• Waivers have been found where the producing party could not describe what they did...so document your procedures.
• Perfection is not expected.

It was pointed out that whether FRE 502 would be binding in state courts, has not been tested.
Citing the Victor Stanley case and Judge Grimm be prepared to answer, "did you do enough to find the privileged documents?"
   
The panel also mentioned the Amobi v. District of Columbia case where Judge Facciola discussed "inadvertent."  The Judge's analogy was paraphrased; "while 502b would allow me to round up the animals and put them back in the barn, were reasonable steps taken to avoid letting the animals out?"
One of the precedents was clear.  When the data had been produced to the expert witness for review, privilege was waived.

The panel then examined what would be considered "reasonable" efforts to find and protect your privileged information.  Keyword search alone is usually not sufficient as it finds only 25% of documents, so you should add in sampling.  The type of vendor you are working with can make a difference.  Do they understand the methodology and work flows around it.

Craig Carpenter provided an extensive list of search features that could help.  He mentioned threading, email de-duping, visualization (being able to see who spoke with whom and when on certain topics), concept search (relate documents that are substantively similar but may not share keywords), clustering-(particular set of keywords-take first few documents that relate to the category), grouping-(more sophisticated clustering-take all documents that relate to the category) automatic categorization (the tool does it for you), amd predictive coding (form of automated review).  Not surprisingly, Recommind's impressive products are capable of helping you with all of these search techniques.   Craig was in "teaching mode" and not in sales mode and his examples and explanations were excellent.


The panel discussed an instance of dealing with 2.5 million documents and the client did not want to pay for privilege review but rather, instructed outside counsel not to turn over anything privileged. LOL!  Obviously it would cost too much.   So, with client's consent, they used automated review tools with some direction from knowledgeable people.  They were able to save 3 or 4 months of processing time and several million dollars.

Mark Michels said "hypothetically," what if he was the "impecunious client?"  "How could they be reasonable and save money too?"  The panel proposed technology search with Bayesian models and sample seed sets to help cull down the data but then made it clear that "at some point, you have to put some eyes on it."  100% automated tools or 100% manual and people-powered processes have not been favored by the courts.

The last "best practice" was to re-iterate that it was critical to define "reasonableness" or, through agreement with the other party(s), take "reasonableness" out of the equation by agreeing what both parties would do for production and what would happen in the event of any inadvertent disclosure.  They closed with one of the best lessons for those who would rely solely on FRE 502 to save you from waiving privilege during disclosure:  Crediting Judge Grimm for the analogy, 502 was "like a bungee cord.  It can save you, but it is still a terrifying experience."