Subscribe
Share
Share
Embed
OK at Work, hosted by Offit Kurman attorneys Russell Berger and Sarah Sawyer, is a weekly podcast that discusses current events and legal issues impacting business owners. From updates on the ever-changing employment law landscape to the risks and benefits of integrating AI into your workplace, subscribe to stay up-to-date on issues and events that may impact you and your business.
Sarah Sawyer: Welcome to this week's
OK at Work with myself, Sarah Sawyer,
my colleague Russell Berger, both
attorneys at Offit Kurman, and today
we are talking about client use of AI.
So our clients often,
obviously search AI for things.
We've mentioned it a couple of times
here and there, in our episodes,
and often sometimes we'll get on
a call with a client and they've
got one of those note takers, and
that's very hard to get rid of.
They show up and the client doesn't
know that they're even there.
And we're like, what's this note taker?
There's a lot of those things going on.
So obviously there's a lot of new
tools that people are using and
are excited to use and often can
be used safely and it can be a good
tool for people, but something that
we're really talking with clients about
paying attention to, is making sure that
they're not destroying attorney-client
privilege and confidentiality,
which is something that, is really
important in the client and attorney
relationship because if it's not
protected and you go, talking to people
about things, you know, so before AI,
there would be the telling
everyone at a bar that you did
something, a problem, right?
So that's not protected, that's not
confidential, that's not privilege.
But when you talk to us about it,
if we're your attorney, it is.
So as a basic concept, there's
a lot more nuance to it.
But now we have folks who are going
and putting things into AI specifically
sometimes open AI, which means that
the terms and service of that open
AI says that they can share it.
Which essentially is like telling
the world whatever you just told AI.
So it's very different than, coming to us.
And there's been some recent cases on this
where clients have gone and used AI prior
to hiring an attorney and then provided
that what they found in AI to their
attorney and then tried to assert that it
was privileged or confidential because now
they've handed it over to the attorney.
And without going too deep into the
weeds, that's not a good practice to
engage in because there's a lot of
reasons that could be susceptible to
being turned over to the other side
in litigation or being produced.
So what are some things, Russell, with
that framework that clients should really
be thinking about as it relates to using
these tools and anything that touches
legal or a case or anything like that.
Russell Berger: Yeah, certainly when
it touches legal and we talked about
this a little bit on an episode we
did a few weeks ago on what to do
when you find how you've gotten sued.
And where things said not to do was
to go run it all through Chat GBT
and, have it spit out your defense.
Yeah.
So don't do that.
I mean the risk there as you mentioned
is that anything your communication with
the AI program is not privileged and
it's out there in a very public way and
maybe there's some arguments in certain
types of terms and conditions, but, it's
much better to, tell your lawyer you want
your lawyer to leverage AI tools to help
flesh out ideas than for
you to do it yourself.
And I think that makes sense anyway
because, if you had a car and it could
only go 25 miles an hour and then we
built an engine that could go a hundred
miles an hour you'd want someone that
was a professional driver at 25 miles
an hour to be the person handling
the wheel at a hundred miles an hour
and it's going four times faster.
It's kinda the same
thing with AI and legal.
It's not especially at this stage,
and maybe this will change in
the not too distant future, but
for now, like AI is imperfect.
It has inaccuracies.
I was at a conference where someone
said, it's right about 80% of the time,
and that's probably gone up a little
bit since I heard that statement.
But I heard 80% of the time think, well,
that's terrible from a legal perspective.
Because it is, and we should strive for
much, much higher accuracy than that.
And that's okay.
You can be right four outta five times,
but that's a pretty high failure rate
when you're talking about legal issues.
And so you want someone who knows how
to operate this high powered fast tool
because, if you don't, it's just as
likely to drive you off the cliff fast.
And you have that accident faster.
Because it's Not more accurate
than, human attorneys at this stage.
And again I think, we could be
doing this in six months, 12
months, a couple years, whatever.
And that might not be the case anymore.
That's certainly where we're headed.
But, I think from a privilege standpoint
and from an accuracy standpoint,
you want to leverage AI in a way, in
conjunction with your attorney, under
your attorney's supervision as opposed
to just going off and doing it on your
own where you haven't read the terms
of the software you're using where
it's not being directed by an attorney.
There isn't, someone that knows the legal
ins and outs who's reviewing the output.
And even if you're gonna put it
into, an AI bot and then hand that
printout to your attorney, I think
we've made this joke before, it'd
be like walking into the doctor's
office with a printout from Web MD.
You don't need to do that.
And you can certainly work with
your attorney to figure out the
best ways to use this new high power
tool, which we should be using.
And we are pro using this
technology, but using it in a safe,
productive privilege kind of way.
Sarah Sawyer: Yeah.
And I think it's exactly to those points.
Anything that's high stakes also and
high something that's important to you
from a legal perspective, you're going to
ultimately end up talking to an attorney.
So start there instead.
As opposed to, 'cause obviously
we have, the handing WebMD is a
great analogy 'cause people come
with something, I've got a fully
baked something, and here you go.
You don't won't have to do anything.
And that's just, almost
never the case, right?
Russell Berger: Yeah.
Sarah Sawyer: And so it,
Russell Berger: it reminds me of
the time I punched in some symptoms
I had in the WebMD and it told me
I had network connectivity issues.
Sarah Sawyer: Love it.
Yeah.
So definitely you're gonna have to
go and talk to the professional.
So just start with that step.
I think that's really the main takeaway.
And I think those are all great
points We're gonna make use of it.
It's a good tool.
But you really have to be very
careful as it relates to these topics.
Russell Berger: Yeah.
That's for sure.
Sarah Sawyer: All right,
well thanks Russell.
We'll see you next time.
Russell Berger: Bye Sarah.