A
new
Disco
study
suggests
that
despite
all
the
hoopla
over
AI,
we
have
a
ways
to
go
before
it
becomes
commonly
used
in
litigation
and,
more
particularly,
eDiscovery.
The
paradox
is
striking:
even
though
Gen
AI
benefits
are
well
recognized,
there
remains
stubborn
reluctance
to
embrace
it.
The
study
is
entitled
Legal
AI:
Driving
the
Future
of
the
Profession.
It
was
primarily
conducted
online
in
the
late
summer
by
Disco
and
Ari
Kaplan
Advisors.
The
underlying
survey
was
completed
by
112
individuals,
about
half
in-house
and
half
in
law
firms.
Interviews
were
also
conducted.
The
focus
was
primarily
on
the
use
of
AI
tools
in
eDiscovery.
Disco
is
a
leading
eDiscovery
provider.
The
Significance
eDiscovery
has
traditionally
served
as
the
proving
ground
for
legal
tech
adoption,
the
proverbial
canary
in
the
coal
mine.
The
fact
is
eDiscovery
is
driven
by
time
pressures
that
are
often
imposed
by
courts
or
clients.
These
pressures
force
lawyers
and
legal
professionals
to
think
about
how
to
get
work
done
quickly
irrespective
of
billable
hours.
When
the
court
orders
you
to
produce
relevant
documents
in
30
days
requiring
you
to
locate
and
review
millions
of
data
sources,
you
don’t
have
time
to
figure
out
how
to
squeeze
out
the
maximum
number
of
billable
hours.
You’re
more
concerned
about
not
being
embarrassed
or
worse.
It
was
these
pressures
that
led
to
things
like
technology
assisted
review
that
sped
up
the
eDiscovery
process
and
begrudgingly
became
the
norm.
Watershed
Moment?
Disco
describes
eDiscovery
as
being
at
a
watershed
moment
with
AI.
Indeed,
there
are
certainly
some
suggestions
in
the
Disco
data
that
a
corner
may
about
to
be
turned
when
it
comes
to
the
use
of
AI.
That
would
make
logical
sense
since
the
efficiencies
and
time
savings
that
AI
tools
could
bring
are
significant.
But
some
of
the
data
suggests
that
while
AI
may
indeed
revolutionize
litigation,
things
are
not
yet
changing
all
that
much.
In
fact,
Disco
itself
concludes
in
its
report
based
on
the
survey,
“Few
—
if
any
—
have
unlocked
its
promise
at
scale.”
Some
Data
Points
For
example,
42%
of
those
in
law
firms
reported
no
external
pressure
to
use
AI
solutions.
This
is
consistent
with
the
findings
of
an
ACC
study,
and
one
done
by
Thompson
Hines,
both
of
which
I
previously
discussed.
Moreover,
36%
of
in-house
attorneys
surveyed
say
they’re
not
facing
pressure
to
use
AI
tools
from
management.
Some
more
key
statistics:
only
35%
of
those
surveyed
report
having
incorporated
GenAI
in
routine
legal
processes
to
any
extent.
That’s
not
terribly
surprising,
since
56%
of
in-house
counsel
say
they
don’t
yet
see
GenAI
as
a
tool
for
controlling
litigation
costs.
The
Paradox
This
reluctance
persists
even
though
70%
recognized
a
top
benefit
of
AI
was
an
increase
in
efficiency.
(Thirty-four
percent
identified
costs
savings
which
amounts
to
the
same
thing).
Fifty-eight
percent
mentioned
better
analysis
and
insights
and
40%
reported
faster
evidence
gathering.
Similar
benefits
like
the
ability
to
quickly
surface
key
evidence,
spotting
patterns
and
themes
from
the
data,
and
the
ability
to
assess
case
merit
earlier
have
been
noted
by
leading
commentators
like
Doug
Austin.
It’s
a
paradox:
while
GenAI
can
make
things
better,
it’s
not
enthusiastically
embraced.
Why?
Why
Indeed?
The
reasons
given
for
not
using
AI
are
a
little
troubling.
Despite
evolving
tools
that
better
ensure
privacy
and
security,
70%
of
law
firm
and
68%
in-house
respondents
said
privacy
and
security
are
still
the
biggest
obstacles.
In
addition,
consistent
with
the
other
surveys
mentioned
above,
both
in-house
and
outside
lawyers
are
concerned
about
demonstrating
ROI.
But
the
main
reason
is
that
legal
professionals
are,
in
my
opinion,
turning
a
blind
eye
to
the
benefits
and
stubbornly
holding
on
to
previous
ways
of
doing
things.
Fifty-two
percent
of
those
in
law
firms,
for
example,
said
they
are
only
using
AI
to
differentiate
their
firm
as
an
“innovator.”
One
lawyer
put
it
this
way,
“Cost
is
not
yet
a
factor;
we
are
less
concerned
with
profit
margin
and
more
on
gaining
market
share.”
Another
said,
“Reduced
costs
are
not
a
reality
yet.”
In
other
words,
law
firms
are
not
using
AI
substantively
or
recognizing
the
benefits.
Several
labeled
the
benefits
of
GenAI
as
“pie-in-the
sky.”
Some
cling
to
the
belief
that
the
failure
rate
with
AI
in
document
review
is
higher
than
with
humans.
Some
said
even
though
blown
away
by
results,
they
would
still
require
a
substantial
amount
of
human
oversight.
This
even
though
79%
of
those
surveyed
rated
GenAI
tools
with
a
three
or
higher
on
a
five-point
scale
when
it
came
to
accuracy
and
53%
said
it
was
a
four
or
five.
Other
Issues
There
are
a
host
of
other
issues
cited
in
the
report
as
rationalizations
for
avoiding
change
(my
comments
to
the
cited
issues
appear
in
parenthesis):
-
Generational
Differences:
There
is
a
belief
that
more
experienced
lawyers
are
more
reluctant
to
use
AI
tools
than
younger
lawyers.
(The
notion
seems
to
be
its
use
is
limited
to
those
who,
due
to
their
limited
experience,
can’t
be
trusted,
and
that
more
experienced
lawyers
won’t
be
able
to
master
it.)
-
Limited
Time:
It
takes
time
—
non-billable
time
—
to
learn
how
to
use
GenAI
tools
(aka
let’s
not
invest
the
time
to
do
things
better).
-
Loss
of
Control:
We
will
lose
control
of
our
data
if
we
put
it
on
an
AI
platform.
(We
heard
that
before
about
the
cloud.
We
know
how
that
turned
out.)
-
Reduced
Billable
Hours:
“speed
will
reduce
revenue.“
(Of
course.)
-
AI
Talent
Deficiency:
There
is
not
enough
AI
talent
available
to
enable
us
to
understand
how
to
use
and
implement
GenAI.
(So
many
commentators,
legal
professionals,
vendors,
and
consultants
are
talking
nonstop
about
AI,
it’s
hard
to
conclude
the
information
and
talent
isn’t
there.)
-
Misaligned
Functionality
Expectations:
GenAI
can
do
some
things,
but
it
just
can’t
solve
most
problems
yet
or
fully
answer
questions.
(Perfect
should
not
be
an
enemy
of
the
good,
particularly
when
the
good
is
recognized.)
-
Accuracy
Concerns:
Accuracy
is
repeatedly
stressed
as
a
reason
to
encourage
continued
human
oversight.
(Accuracy
is
a
concern
but
that
doesn’t
mean
throwing
the
baby
out
with
the
bath
water.)
A
Silver
Lining
But
there
may
be
a
silver
lining
when
it
comes
to
AI
and
eDiscovery:
if
history
is
any
guide,
things
may
eventually
change.
As
I
said
at
the
outset,
eDiscovery
is
the
canary
in
the
coal
mine
when
it
comes
to
technology
adoption
by
the
legal
profession.
All
too
often,
the
pressure
to
get
eDiscovery
work
done
quickly
trumps
the
desire
to
resist
change.
We
have
seen
efficiency
tools
like
technology
assisted
review
and
continuous
machine
learning
gradually
become
well
accepted
and
standard,
particularly
as
data
and
data
sources
exploded
exponentially.
It
did
take
years
to
overcome
the
notion
that
humans
had
to
do
everything,
but
we
did
get
there.
The
New
Reality
So,
it’s
tempting
to
conclude
that
despite
all
its
benefits,
many
of
which
are
already
recognized,
the
adoption
of
GenAI
will
take
the
same
slow
and
torturous
course.
But
there
is
another
reality
revealed
by
the
survey:
96%
say
eDiscovery
workloads
are
increasing
or
staying
the
same.
And
there
is
a
recognition
that
a
variety
of
new
data
sources
including
prompts
and
outputs
must
be
dealt
with.
Fifty-two
percent
of
those
surveyed
say
these
new
sources
will
make
the
litigation
cycle
longer,
inevitably
increasing
costs.
These
new
data
sources
and
continued
time
pressures
may
force
legal
professionals
to
adopt
GenAI
tools
out
of
necessity
just
to
keep
up
and
satisfy
the
demands
of
courts,
regulators,
and
clients
—
just
as
they
gradually
did
with
the
adoption
of
TAR,
only
quicker.
eDiscovery
Is
Cool
That’s
why
watching
what
happens
in
eDiscovery
is
important:
legal
will
be
forced,
even
though
kicking
and
dragging
their
feet,
to
adopt
GenAI.
Time
constraints
and
risk
aversion
may
force
the
adoption
many
are
currently
avoiding.
Or
as
one
respondent
put
it,
perhaps
a
bit
reluctantly,
“AI
has
made
eDiscovery
cool.”
Stephen
Embry
is
a
lawyer,
speaker,
blogger,
and
writer.
He
publishes TechLaw
Crossroads,
a
blog
devoted
to
the
examination
of
the
tension
between
technology,
the
law,
and
the
practice
of
law.