When
Kara
Peterson
co-founded
descrybe.ai,
she
wasn’t
just
launching
a
legal
tech
startup.
She
was
applying
the
logic
of
public
health
to
the
legal
system.
She
saw
something
lawyers
often
miss:
that
legal
information,
like
clean
water
or
vaccines,
becomes
life-altering
only
when
it
is
accessible,
understandable,
and
widely
distributed.
Kara
is
not
a
lawyer.
Her
background
is
in
public
health
and
communications.
Yet
she
now
leads
a
company
that
has
used
AI
to
summarize
more
than
3.3
million
judicial
opinions
and
made
them
free
and
publicly
available.
In
our
conversation
on
“Notes
to
My
(Legal)
Self,”
she
explained
how
justice,
like
health,
is
a
public
good.
And
access
to
law
is
one
of
its
critical
delivery
systems.
Legal
Transparency
Is
A
Health
Equity
Issue
Kara
grew
up
in
Madison,
Wisconsin,
in
a
culture
steeped
in
social
justice.
Her
early
experiences
shaped
a
worldview
that
makes
her
uniquely
suited
to
build
bridges
between
legal
complexity
and
everyday
reality.
As
she
puts
it,
“Access
to
the
law
is
so
deeply
connected
to
other
systemic
problems.
This
is
a
way
to
attack
something
that’s
been
broken
for
generations.”
From
a
public
health
lens,
the
connection
is
obvious.
People
struggle
to
find
housing,
defend
themselves
at
work,
or
navigate
family
law
—
not
because
they
lack
intelligence
but
because
they
lack
access.
And
when
legal
literacy
is
only
triggered
by
trauma,
as
when
someone
is
fired,
evicted,
or
sued,
we
are
setting
up
millions
of
people
to
fail.
Legal
information,
when
made
legible
before
the
crisis
hits,
becomes
preventive
infrastructure.
It
is
no
different
than
distributing
hand-washing
guides
or
vaccination
instructions.
Human
Bias
Already
Exists.
AI
Just
Makes
It
Visible
Much
of
the
fear
surrounding
AI
in
law
is
valid.
But
Kara
is
quick
to
point
out
that
many
of
the
risks
we
associate
with
machines
are
already
deeply
embedded
in
human
systems.
“If
we
had
a
judicial
system
that
was
completely
fair
and
didn’t
ever
take
people’s
biases
into
account,
that
would
be
hard
to
match.
But
we
know
that’s
not
how
it
works.”
Her
point
is
not
that
AI
is
perfect.
It
is
that
our
current
legal
infrastructure
is
far
from
it.
And
if
we
judge
new
tools
by
perfection
while
accepting
the
flaws
of
legacy
systems
as
inevitable,
we
stall
innovation.
Instead,
she
suggests
we
approach
AI
as
we
would
a
new
public
health
tool.
We
evaluate
its
safety.
We
understand
its
limits.
We
measure
impact.
And
we
keep
a
human
in
the
loop
not
to
preserve
tradition,
but
to
preserve
context
and
accountability.
Law
Should
Not
Be
A
Trauma-Only
Language
One
of
the
most
striking
moments
in
our
conversation
came
when
we
discussed
how
most
people
“learn”
the
law.
We
don’t
learn
employment
law
until
we’re
fired.
We
don’t
understand
custody
law
until
we
divorce.
We
don’t
engage
with
IP
law
until
our
work
is
stolen.
“It’s
always
after
something
painful,”
Kara
noted.
“And
even
when
you’re
educated,
even
when
you
have
attorneys
in
the
family,
it’s
still
incredibly
hard
to
understand
the
system.”
This
is
exactly
where
legal
tech
can
intervene,
not
just
to
reduce
time
or
cost,
but
to
reduce
the
harm
of
delayed
understanding.
It’s
why
at
TermScout,
we
certify
contracts
not
just
for
fairness,
but
for
clarity.
We
want
contract
reviews
to
feel
less
like
damage
control
and
more
like
informed
consent.
When
people
can
see,
compare,
and
understand
the
terms
of
a
deal
in
plain
language,
they
gain
agency.
Justice
Tech
Is
Business
Tech
For
The
Rest
Of
Us
What
Kara
and
her
team
at
descrybe.ai
are
doing
parallels
what
we
do
at
TermScout.
They’re
making
judicial
opinions
legible.
We’re
making
contracts
legible.
Both
models
are
designed
to
scale
clarity,
not
just
output.
And
that’s
the
deeper
takeaway
here.
The
justice
gap
and
the
contract
gap
are
part
of
the
same
problem:
society
has
accepted
legal
opacity
as
normal.
But
it
doesn’t
have
to
be.
If
we
can
explain
a
30-page
legal
opinion
to
a
nonlawyer,
we
can
explain
a
vendor
contract
to
a
startup
founder.
If
we
can
summarize
a
complex
ruling
with
an
AI
model,
we
can
certify
a
set
of
terms
with
a
standard.
Kara
reminded
us
that
not
all
impact
lives
in
the
courtroom.
Some
of
the
most
meaningful
change
comes
when
you
give
people
tools
before
they’re
in
crisis.
The
law
is
not
just
for
the
few.
It
is
infrastructure.
And
when
we
treat
access
to
it
like
a
public
health
mission,
we
stop
asking
how
much
harm
we
can
tolerate
and
start
designing
systems
that
prevent
it
altogether.
Olga
V.
Mack is
the
CEO
of TermScout,
an
AI-powered
contract
certification
platform
that
accelerates
revenue
and
eliminates
friction
by
certifying
contracts
as
fair,
balanced,
and
market-ready.
A
serial
CEO
and
legal
tech
executive,
she
previously
led
a
company
through
a
successful
acquisition
by
LexisNexis.
Olga
is
also
a Fellow
at
CodeX,
The
Stanford
Center
for
Legal
Informatics,
and
the
Generative
AI
Editor
at
law.MIT.
She
is
a
visionary
executive
reshaping
how
we
law—how
legal
systems
are
built,
experienced,
and
trusted.
Olga teaches
at
Berkeley
Law,
lectures
widely,
and
advises
companies
of
all
sizes,
as
well
as
boards
and
institutions.
An
award-winning
general
counsel
turned
builder,
she
also
leads
early-stage
ventures
including Virtual
Gabby
(Better
Parenting
Plan), Product
Law
Hub, ESI
Flow,
and Notes
to
My
(Legal)
Self,
each
rethinking
the
practice
and
business
of
law
through
technology,
data,
and
human-centered
design.
She
has
authored The
Rise
of
Product
Lawyers, Legal
Operations
in
the
Age
of
AI
and
Data, Blockchain
Value,
and Get
on
Board,
with Visual
IQ
for
Lawyers (ABA)
forthcoming.
Olga
is
a
6x
TEDx
speaker
and
has
been
recognized
as
a
Silicon
Valley
Woman
of
Influence
and
an
ABA
Woman
in
Legal
Tech.
Her
work
reimagines
people’s
relationship
with
law—making
it
more
accessible,
inclusive,
data-driven,
and
aligned
with
how
the
world
actually
works.
She
is
also
the
host
of
the
Notes
to
My
(Legal)
Self
podcast
(streaming
on Spotify, Apple
Podcasts,
and YouTube),
and
her
insights
regularly
appear
in
Forbes,
Bloomberg
Law,
Newsweek,
VentureBeat,
ACC
Docket,
and
Above
the
Law.
She
earned
her
B.A.
and
J.D.
from
UC
Berkeley.
Follow
her
on LinkedIn and
X
@olgavmack.