2 + 2 do e sn t a lwa ys = 4 (re me mb e r we a re ta lking a b - PowerPoint PPT Presentation

 2 + 2 do e sn’ t a lwa ys = 4 (re me mb e r – we a re ta lking a b o ut la wye rs)  Cla sh o f diffe re nt c ulture s with diffe re nt o b je c tive s  A wa y fo rwa rd 1

2

 Bla ir & Ma ro n (1985)  T he Se do na Co nfe re nc e , T he Se do na Co nfe re nc e Be st Pra c tic e s Co mme nta ry o n the Use o f Se a rc h a nd I nfo rma tio n Re trie va l Me tho ds in E -Disc o ve ry (2007)  T RE C L e g a l T ra c k (NI ST )  Re c e nt c o mme nta rie s b y Ba ro n, Oa rd, Gro ssma n, Co rma c k, a nd o the rs 3

 Disability Rights Co unc il o f Gr e ate r Wash. v. Wash. Me tr o . ansit Auth. (2007 ) T r › Judg e F a c c io la re c o g nize s c o nc e pt se a rc hing , a s o ppo se d to ke ywo rd se a rc hing , “is mo re e ffic ie nt a nd mo re like ly to pro duc e the mo st c o mpre he nsive re sults.”  Vic to r Stanle y (2008) › Judg e Grimm po ints to g ro wing b o dy o f lite ra ture tha t hig hlig hts the risks o f c o nduc ting a n unre lia b le o r ina de q ua te ke ywo rd se a rc h o r re lying e xc lusive ly o n suc h se a rc he s.  Nat’l Day L abo r e r Or g. Ne two r k v. U.S. Immig. & Custo ms c e me nt Age nc y (2012 ) E nfo r › (Sc he indlin) “pa rtie s c a n (a nd fre q ue ntly sho uld) re ly o n la te nt se ma ntic inde xing , sta tistic a l pro b a b ility mo de ls, a nd ma c hine le a rning to o ls.” 4

5

• Da Silva Mo o re v. Pub lic is Gro upe • I n re Ac to s Pro duc ts L ia b ility L itig a tio n • Glo b a l Ae ro spa c e , I nc . v. L a ndo w Avia tio n, L P • K le e n Pro duc ts L L C v. Pa c ka g ing Co rpo ra tio n o f Ame ric a • I n re Bio me t M2a Ma g num Hip I mpla nt Pro duc ts L ia b ility L itig a tio n 6

 Da Silva Mo o r e (F e b. 2012) › F irst o pinio n e ndo rsing T AR › Issue d b y no te d jurist (Pe c k) › Co mpute r a ssiste d re vie w is a c c e pta b le wa y to se a rc h fo r re le va nt E SI in a ppro pria te c a se s. T ra nspa re nc y de e me d vita l: de fe nda nts re q uire d to turn o ve r the ir › se e d se t to pla intiffs, inc luding do c ume nts ta g g e d a s no n-re spo nsive in the tra ining ro unds, to e na b le pla intiffs to de te rmine whe the r the c o mpute r wa s b e ing a ppro pria te ly tra ine d.  Glo bal Ae r o spac e (Apr il 2012) › Sta te c o urt c a se › Pe rmits de fe nda nts, o ve r pla intiffs’ o b je c tio n to use T AR me tho do lo g y to se a rc h fo r re spo nsive do c ume nts › Co urt did no t spe c ific a lly e ndo rse the te c hno lo g y o r c o nc lude it wa s b e tte r suite d tha n o the r le g a l se a rc h to o ls › Re c o g nize d tha t pla intiffs still ha d rig ht to o b je c t po st-pro duc tio n 7

 Ac to s CMO issue d in July 2012 › Custo dia ns use d fo r sa mple c o lle c tio n › Size o f ra ndo m c o ntro l se t › E a rly pro duc tio n o f privile g e lo g › Pa rtie s no mina te six e xpe rts to re vie w c o ntro l se t › F o llo wing re vie w o f c o ntro l se t using a c tive le a rning mo de l, e xpe rts wo rk c o lla b o ra tive ly to de te rmine re spo nsive ne ss › Co urt ma nda te s suffic ie nt tra ining ro unds › Pa rtie s me e t a nd c o nfe r to a g re e o n re le va nc e sc o re 8

o duc ts a nd Bio me t fo llo w re a so ning in  Kle e n Pr Glo b a l Ae ro spa c e  Kle e n Pr o duc ts › Pla intiffs a rg uing fo r use o f T AR a s do -o ve r to e a rlie r a pplic a tio n o f ke ywo rd se a rc h me tho do lo g y › Pla intiffs la te r a g re e d to withdra w de ma nd fo r e xisting pro duc tio n re q ue sts  Bio me t › Co urt re fuse s to disturb de fe nda nts’ unila te ra l de c isio n re g a rding pro c e ss use d fo r T AR se a rc h a nd c ulling 9

10

• Sc o pe o f disc o ve ry unde r F RCP do e s no t inc lude “disc o ve ry a b o ut disc o ve ry” • Disc lo sure o f pro te c te d wo rk pro duc t • Unne c e ssa ry risk o f a dditio na l litig a tio n whe n no n-re spo nsive info rma tio n disc lo se d • Atto rne y c e rtific a tio n o f c o nduc t o f re a so na b le se a rc h fo r re spo nsive do c ume nts sho uld suffic e 11

 I g no ra nc e  Co mfo rt with o lde r me tho do lo g y  E xpe nse  T o da te , o nly a fe w judic ia l o pinio ns ha ve a ddre sse d T AR  F e a r o f “do -o ve r” if Co urt do e s no t e ndo rse me tho do lo g y 12

13

• Co lla b o ra tive • Se e k pe rfe c tio n a b o ut se a rc h re sults • I nte re ste d in disc lo sure ra the r tha n de fe nsib ility • L e ss c o nc e rne d a b o ut c o sts? 14

• Se e k de fe nsib ility; no t pe rfe c tio n • T rying to win; no t inte re ste d in he lping o ppo sing pa rty • Re stric te d fro m divulg ing privile g e d info rma tio n • Wo rrie d a b o ut the b o tto m line • T ra ine d no t to ta ke risks • Will a do pt a dva nc e d se a rc h me tho do lo g ie s if in b e st inte re st o f c lie nt o r if o rde re d b y c o urt 15

16

 Po te ntia l disc lo sure s: › Spe c ific so ftwa re › Na ture o f do c ume nts › Size o f c o rpus › Culling te c hniq ue s to b e e mplo ye d › Use o f c o ntro l se t › Sa mpling te c hniq ue s use d to g e ne ra te “se e d se t” › Wo rkflo w/ pro c e ss › E xpe rie nc e / e xpe rtise o f re vie we rs › Re le va nc e / re spo nsive ne ss thre sho lds › Pre c isio n/ re c a ll ra te s 17