August 20, 2025 • 40 mins
Comment on the Show by Sending Mark a Text Message.
Artificial intelligence has revolutionized how we approach many aspects of life, including legal analysis, but what happens when employees rely on AI to evaluate potential employment discrimination cases? This episode uncovers the dangerous pitfalls that can trap unwary workers seeking justice.
The allure of immediate feedback from AI tools like ChatGPT has led many employees to trust these systems with complex legal analysis. Mark explains why this approach often fails – from AI's inability to access crucial case law behind legal database paywalls to the shocking reality that these systems sometimes fabricate non-existent court cases and legal principles. He shares real examples of lawyers who were sanctioned by federal judges after submitting AI-generated research containing completely fictional legal citations.
Beyond accuracy issues, using AI for legal analysis raises serious confidentiality concerns. When you input sensitive workplace details into public AI systems, you may inadvertently violate employment agreements and expose private information. Mark provides practical guidance on how to anonymize your narrative while still getting useful insights.
What makes this episode particularly valuable is the actionable framework Mark provides for conducting effective legal research. He walks listeners through a step-by-step process that combines AI's strengths with traditional legal research methods – writing a detailed chronological narrative, using AI to identify relevant legal standards, then verifying and deepening this understanding through researching actual court decisions in your jurisdiction.
Most compelling is Mark's empowering message that the legal system, while imperfect, remains accessible to employees willing to invest time in understanding their rights. He shares examples of individuals who successfully negotiated settlements without attorney representation by presenting well-documented, legally-informed narratives of their experiences.
Whether you're currently facing workplace discrimination or simply want to understand your rights better, this episode provides crucial knowledge about how to use AI as one tool in your legal arsenal – while recognizing when human legal expertise remains irreplaceable. Subscribe to the Employee Survival Guide for more insights that help level the playing field between employers and employees.
If you enjoyed this episode of the Employee Survival Guide please like us on Facebook, Twitter and LinkedIn. We would really appreciate if you could leave a review of this podcast on your favorite podcast player such as Apple Podcasts. Leaving a review will inform other listeners you found the content on this podcast is important in the area of employment law in the United States.
For more information, please contact our employment attorneys at Carey & Associates, P.C. at 203-255-4150, www.capclaw.com.
Disclaimer: For educational use only, not intended to be legal advice.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:01):
Hey, it's Mark and welcome back to another edition
of the Employee Survival Guide.
Today's topic is the hiddendangers of using AI to predict
your employment discriminationcase.
In today's world, employees canturn to artificial intelligence
AI for almost anything, evenlegal advice.
Some employees are beginning touse AI tools to quote-quote
(00:22):
predict whether they have enoughstrong employment
discrimination case.
Employees are tempted with theimmediate feedback that AI
provides in analyzing theirlegal cases.
We have recently seen employeeswho conducted their own legal
research and put together whatthey believe are the correct
legal citations mostly statutesand prominent US Supreme Court
(00:43):
decisions, along with their ownlegal analysis as to why they
have a case of employmentdiscrimination and retaliation.
This is their pitch toemployment lawyers like myself
to review their cases, or theiractual or they're actually
filing it with the state andfederal and city agencies
regarding their claims ofemployment discrimination,
(01:04):
hostile work environment andretaliation.
This came up, this topic came upbecause I receive quite a few
of these type of emails wherepeople are trying to pitch a
story to me and they, you know,no fault of their own.
I mean, ai is out there.
It's new and novel andemployers are using it, of
(01:28):
course, but you know, peoplehave access to a Google notebook
or a chat, gdp or something,and the technology itself is
relatively new and it's basedupon you know, upon what you fed
into the system.
And so when I read these emailsand they give me the cut and
(01:50):
paste of what their research wasand we all know what this looks
like it's like this chat orthis AI type of production of
work product that it looks likeit's usually a list or that's
kind of stale, it's not veryemotional, it's just a matter,
or that's kind of stale, it'snot very emotional, it's just
matter of fact, whatever.
But it's only based upon whatthe AI device has captured in
(02:11):
its learning.
And the funny thing about thistopic is that the AI devices,
these algorithms, can pick upinformation related to agency
websites and devour that, butthe case law material, the case
decisions from our courts statecourts, federal courts they're
(02:33):
not readily accessible andthey're behind paywalls where
you have to go in and access it.
Now there are cases reportedout there that there are various
websites that put it out, butthe vast majority of case
decisions are locked behindpaywalls for websites called
like Westlaw and LexisNexis, andso the AI devices that the
(02:57):
employees are accessing don'thave that wealth of depth so
that information is missing fromits analysis.
Now, this is a technology ininfancy, so you can imagine that
if employees are puttingtogether their case analysis
today, it might get more robustas the technology increases,
(03:18):
maybe what they feed into it.
I would love to see the AIdevices like a chat GDP, have
all of the case materialdecisions going back, let's say,
just give it 10 years.
I mean these devices can learnthe machine learned very, very
quickly device so the employeescould access it, because the
(03:45):
employees would basically beable to read public court
decisions because courts arepublic and read these decisions
and decide for themselves ifthey have a case or not.
Remember, lawyers are not theonly parties that have an angle
or the access to the law.
Everybody has access to the lawand at least that's my opinion
(04:07):
and that's why I try to put theslant I do on these podcast
episodes.
I give you case decisions.
You can hear about it.
So the so having employeesaccess in case decisions is
really uber important toleveling the playing field
(04:29):
between employers and employees,which is not even, which is
uneven, the whole at-will debateyou hear about me raise all the
time.
So back to the point ofemployees trying to figure out
their cases before they confirma lawyer or file before an
agency and they're doing theirAI research.
I'm not saying don't do this.
I'm saying just do this withthe eyes wide open about what
(04:51):
you're looking at, and it's okayto begin to learn like a lawyer
, and this podcast is reallydesigned to kind of tell you
some pitfalls, but also tell youhow do you approach it to do it
better, but also tell you youknow how do you approach it to
do it better.
You know how do you take yourAI research and create it in
such a way that it's moreaccurate than it wasn't.
(05:13):
So first is you know, read Igot a lot of material on my
website and a lot of podcaststalk about it.
For example, hostile at WorkEnvironment Everybody loves that
topic on my website and thepodcast, and so there's a lot of
information about it.
(05:33):
For example, hostile workenvironment Everybody loves that
topic on my website and in thepodcast, and so there's a lot of
information about that.
But how do you interpret ahostile work environment with
respect to the you know, a factpattern that you, what you're
currently going through.
So what I encourage you to dois this why don't you write out
your fact pattern and put it inlong form you know your
narrative from A to Z,chronological order, what
happened, what happened next,who knew about it, whatever, and
write it out?
(05:54):
You might want to not includenames of the employer and use
fictional names of characters,because you're putting this
information publicly out thereinto a domain AI device that's
learning it and you haveconfidentiality provisions of
your when you sign up foremployment, so you would be
careful what you're putting outthere.
(06:14):
I mean, it's like the blackhole, this AI device.
So just caution.
You Just write a narrative offacts and then put it into the
algorithm, the chat GDP, and askit to assess based upon, let's
say, you live in New York Cityand you say AI chat GDP, why
(06:34):
don't you assess it under NewYork City law, human rights law,
new York State human rights lawand Title VII, 1964 Civil
Rights Act in New York State,and let's see how that chat GDP
device pushes out the outcome.
Now we can't trust what it'sgoing to say, but hopefully, if
(06:57):
you're using your fact pattern,you might get a closer vantage
point to illegality, toillegality.
Sometimes when I do the searcheson chat GDP.
I'll get warnings that they'renot a lawyer and they can't make
legal or give legal advice.
You might see that as well, butplay around with it and you
might be able to come very closeto, let's say, your fact
(07:18):
pattern, directing you to theconclusion that you have a
hostile work environment.
And everybody knows what ahostile work environment is.
It's, you know, a pervasive useof discriminatory comments or,
you know, comparison of youversus others that are given
you've received more favorabletreatment.
It's typically in a sexualconnotation or sometimes it's in
(07:41):
a racial connotation where youhave, like you know, pictures of
nooses for a racial case orlightning bolts for the KKK or
some pervasive hostile likeextreme situation which is, and
maybe the computer can spit outthat assessment for you.
(08:01):
So go through that exercise,begin to learn more.
If you spend the time, you canfigure out that there's
information out there and I'mgoing to tell you this.
The next chat GDP search shouldbe give me notable let's say
Southern District of New Yorkcases in the last five years
(08:23):
involving hostile workenvironment related to sex and
see what those decisions come upwith.
And let's say you get five, sixdecisions and you push those
case names back into a Googlesearch and see if you can come
across like a Cornell decision.
They have a website that hascase decisions.
Justia is another site that has, you know, pdfs.
(08:46):
You can read the actualdecision, but use the devices,
the search, in a way that youcan use ChatGP to identify cases
.
You can use a variety ofsources to identify cases on
sexual or hostile environment inthe Southern District of New
York.
It's a federal court and see ifyou can find the cases
themselves to read through themand see if you can find the
(09:07):
cases themselves to read throughthem.
What you're trying to do here isread through the narrative of
fact in the decision, but thenalso look at the court's
analysis of that, and that'swhere I really want to bring
your attention.
What's wrong with the AI use byemployees today is that the
employees don't have anyunderstanding of the analysis as
(09:27):
it relates to being a lawyerand giving the assessment.
Now, I'm not saying thatemployees have to become quick,
off-the-cuff attorneys toidentify their cases.
It's much simpler than that.
I think you have the ability tolook at here's what a case
decision is from a judge andsays here's the law I'm applying
to this narrative of this casedecision you just found using
(09:50):
research, and you can see howthe judge moves through that
analysis, using the fact andthen applying the law to the
fact to establish the conclusion.
Well, there you have it.
You've just gone to law school.
You've learned what the factsare, you've learned the law part
of it and you've learned theanalysis fact.
In law that's called issuespotting or I don't want to give
(10:13):
anything to law school.
It's just looking at facts,looking at law, seeing how the
facts apply and how the law mayapply, and it's a very generic
way of looking at the strength,weakness of a case.
So it's a very generic way oflooking at the strength,
weakness of a case.
Now you would have to see a lotmore of those cases to decide
well, do I have a case or not,or is it how strong it is?
I think you're better offwriting a fact pattern and using
(10:37):
your research, saying, ok, Ifound a sexual harassment case
with a hostile work analysis byJudge so-and-so, and I put that
into my analysis and maybe I'mgoing to search further because
I want to find some more cases.
And here's a trick Once I startto find a commonality of case
decisions about the same thingin different court cases in the
(11:01):
same district, let's say theSouthern District of New York,
which I'm admitted to.
I stop and I realize I have thebasic foundation of what sexual
harassment, hostile workenvironment is for the Southern
District and you want to movefurther into the analysis and
you look at possible otherclaims that might exist and use
(11:21):
the case decision you're readingor researching to identify more
cases.
Now I'll give you a cheat okay,employment discrimination cases
fall into the protectedcategory.
Okay, and it's age, sex,whatever race, religion,
(11:41):
national origin, disability, andyou know, if you're, by and
large, you're going to fall intoone of those buckets or
retaliation or whistleblowing,and you want to do that search
by those different categories ofdiscrimination to find cases
that are applicable to yours.
Now, you're not going to findthe exact case applicable on
point.
They call it.
That takes a little bit deeperdive into the case law, spending
hours trying to locate cases,but you're going to get a quick
(12:03):
assessment using AI, usingGoogle search, using Justia, the
website, to locate cases toread.
You have to read this stuff.
Don't expect ChatGDP to push outthe analysis for you like we're
used to, because that's not theway it works.
What are you missing?
You're missing the subtletiesthat the law is.
You're missing the subtletiesto which an employment lawyer
(12:25):
will give you.
But I think you can come close.
I think you can educateyourself about the laws, because
I've seen clients do thisbefore AI got involved.
I've seen clients go and do theresearch at law schools and law
libraries and they came prettyclose to what the actual
analysis is.
So there's a lot of commonalityof case claims that you can pick
(12:50):
up very quickly to put in.
I know that ChatGDP will pushout.
Here's what sexual or hostilework environment will mean, and
you'll start to.
You know, because you put it inseveral times, you get the same
answer.
That's not the case.
If you actually read a courtdecision.
By doing your deep diveresearch, you can actually
figure out how the courts arelaying out the elements of the
claim for hostile workenvironment.
(13:13):
Then the next stage of that iswell, step back and look at your
fact pattern in the analysisand see, you know, are there
inferences in your fact patternthat support the possibility
that you have a claim of hostileworkment discrimination?
We're looking for stronginferences.
So we're going to take all.
(13:34):
There's two types of evidencehere.
There's direct statements,statements we don't like you
because you know you're an f andwhatever, and uh, and they
start to do extreme things inthe office and there's uh, just
hostility, whatever, uh, andthen there's that's all direct
statement, stuff from a boss, uh.
And then you have this andthat's like five percent of the
cases.
So it's out there but it'ssmoking on the evidence and it's
(13:56):
not really what you'retargeting.
You're targeting circumstantialevidence.
It's how you're treateddifferently than other people
who are treated more favorably,who are the different categories
.
So, if it's sex, hostile workenvironment you're looking at,
and you're female, you'relooking at someone who's a male,
who's treated better, and youwant to do this compare and
contrast, assessment in your ownfact pattern, looking for
(14:19):
inferences of how they'retreated differently.
Assessment in your own factpattern, looking for inferences
of how they're treateddifferently.
I mean, that's really where thepart of I want you to pick up
on.
You know, look at chat GDP, getyour data, your information
about what the law is, do someadditional research to confirm
you're correct, because AI isnot perfect yet.
Get an understanding of workingknowledge of what hostile work
environment is, look at yourfact pattern and then begin to
(14:42):
do your analysis, I think fivethings stand out.
Five circumstantial pieces ofevidence point out of compare
and contrast, and I was firedbecause I didn't receive the
same treatment they did.
Let's not make it complicated.
Okay, put your facts together.
Do your research.
What's going to happen to youis when you do your research and
you spend the time reading,like lawyers do in these cases,
(15:04):
or listening to a podcast abouthostile work environment on my
podcast, you're going to beginto become familiar, like with
anything, the subject matterthat you're working with, and
you'll start to see the nuancesof things.
You'll start to see things popup for you factually that you
may want to put back into yournarrative.
Spend some time.
Don't just shotgun the approachof saying here's a stretch GDP,
(15:26):
I'm going to put my informationand, bang, I'm going to send
that to the lawyer.
Okay, I'm not kidding, thatdoes happen.
It irritates me.
What I'm trying to say is youcan figure out and come pretty
darn close to what your case isif you go about this process of
assessing what lawyers do.
They ask people to write thefact pattern, write the
(15:47):
narrative chronologically.
Don't provide conclusions, juststate the facts, ma'am, and
then do some research andunderstanding what this is?
Because if you're going to aska lawyer to take a case on
contingency which is even worseask a lawyer.
If you're going to ask a lawyerto take a case on contingency
which is even worse you want tobe convincing.
And if you're not convincing inyour initial you know email,
(16:12):
it's going to take about twoseconds for me to realize it and
I'm going to basically dispensewith the.
You know I'm not going to writeback to you, I'm just going to
say that's not a case interestspending my time on.
If you're going to hire a lawyeron an hourly basis and you're
throwing your own money at it,you want to know that you're
hiring an employment lawyer andyou're spending your money
wisely.
There you become acollaborative effort with the
(16:34):
employment attorney and I'veseen clients who are bright and
they're pretty savvy and theycan.
They work in a kind of acollaborative effort to push
along a case and they understandit and so you can age yourself
in the process of understandingthe legal services you're
receiving and understand yourcase.
(16:55):
But if you go about this aspectof the strategy of pushing
things into chat GDP, it's goingto be met with like false left
and right red flags.
People pick it up prettyquickly.
Employers and their counsel canread it very quickly, say oh,
that was obviously AI produced,et cetera.
So people are doing that.
(17:15):
But here's the things to avoid.
To get down to the nitty grittyand it's you know, when you're
creating your narrative andyou're doing your research and
you're going the extra mile tofigure out what it is and it is
not sexual harassment, hostileenvironment.
Beware of the following issues.
Okay, ai oversimplifies the law.
Implement discrimination iscomplex, fact-specific and
(17:36):
constantly evolving.
Complex, fact-specific andconstantly evolving.
Now I told you kind of thebasic primer of how to look at
these cases in terms of yourclaims, because I'm doing that
deliberately, I want to dumb itdown, because I want to make it
more accessible to you.
The law has a tendency to makeit really complex, but I'm
trying to make it easier toabsorb and there are a lot of
(18:01):
commonalities amongst claims andyou can only see that when you
look at it long enoughexperiencing it.
So AI tends to be applied broad,rigid rules, missing exceptions
or subtleties that I wouldunderstand as an employment
lawyer, and they are exceptionsthat might make or break your
case.
I'll give you an example.
(18:24):
So you have a claim.
You are 365 days, 366 days,since you were fired, and it's
well, and you decide finally todo something about it.
What you don't know is thatwhen you filed your UOC charge,
you were a little bit latebecause the statute of
limitations says 300 days, soyou wouldn't know is that when
you filed your USC charge, youwere a little bit late because
the statute of limitations says300 days, so you wouldn't know
(18:45):
that.
So that's an exception.
So just AI oversimplifies thelaw.
So be very, very careful.
Next thing is wrong predictions, wrong decisions.
I start this off with the famousset of cases that came out of
the Southern District of NewYork.
This involved lawyers.
(19:05):
It's a true story.
It was reported in the pressTwo cases of lawyers who were
before the federal judges in twoseparate cases.
Both had conducted AI researchon legal cases, had conducted AI
research on legal cases and thejudge got a little bit upset in
(19:27):
both cases and said and theywere sanctioned because the AI
that they were using whateverthat was had made up case names
and made up case decisionalanalysis, meaning made it up.
I don't know what they wereusing, but both lawyers in both
cases got reprimanded.
And it's very embarrassing.
The judges look at every singlecitation that we give to them
(19:49):
because we're looking at thecases and saying this is
supportive of our client's case.
So, likewise, I'm asking you todo the same thing when you do a
research through ChatGDP Followup and read the case.
You know these lawyers in thatcase that I'm citing the two
cases they did not read thedecisions that this research
chatbot had created for them.
(20:09):
It's embarrassing, but itproves the point that even
lawyers could do the samescrew-up mistake.
So don't do that.
So that's an example ofnon-existing cases and
non-existing case decisionanalysis.
So AI isn't trained on yourunique facts or the latest court
(20:30):
rulings.
That means it can give falseconfidence quote I will win the
case or unnecessarydiscouragement.
I don't have a case leading youto make damaging choices.
So AI is in its infancy.
Like I explained before, itdoesn't have all the thousands
and hundreds of thousands ofcase decisions across the
(20:51):
country in its data bank becauseit's beyond a paywall.
Yeah, you're like.
Well, that's unfair.
Why is that?
Courts are open.
Why can't we access it?
Well, it's a larger question.
I think that we should haveaccess to all the case decisions
in both state and federalcourts.
But maybe there's a reason whythat's not the case, right,
(21:11):
because more information ispower the more power you have,
the more damage you can do on acorporation, so maybe that's the
case.
There's also Westlaw and Lexiswho serve this services, and I
pay a good dollar for thatservice my lawyers myself.
The next item for what AI is notto be trusted is missed
(21:31):
deadlines and lost rights, likethe example I gave you.
If you come up with your hey, Iwant to file a new EOC case 365
days after the adverse actionmeaning you got fired you're too
late.
365 days after the adverseaction meaning you got fired,
you're too late.
You got to file your decisionor your claim to exhaust your
administrative remedies in 300days or within 300 days, so
easiest is missed claims becauseyou had.
(21:54):
You didn't understand thefiling requirements in terms of
deadlines.
I'll give you a basic primer180 days and 300 days on
employment discrimination cases.
Okay, let's just follow that.
So the adverse actions have tooccur within that.
Most state and federal agencieswork on what's called a
work-sharing agreement and soyou get the benefit of a 300-day
window, and I would alwaysencourage you to file with the
(22:17):
EOC first and then have the EOCfile the claim with the local
state agency.
So you get the 300-day windowtime period.
The next item I spoke aboutbefore is when you create your
written narrative and you put itinto the chat GDP to figure it
out.
Use generic names.
(22:40):
Don't use company informationthat's identifying of the
company.
When you type your story into apublic AI system, you expose
sensitive details names, dates,medical history that you can't
take it back.
Ai doesn't guarantee a privacyand you sign most likely
confidentiality agreements withyour employers not to share the
(23:04):
information publicly.
You're putting the informationin a chat GDP or an AI device.
You have disseminatedinformation out there publicly.
So be very careful because it'snot about if.
It's about when these systemsand employers catch up that they
can figure out who disclosedthe information.
So be generic with your namesand company information and what
(23:26):
you're talking about, so itcan't be disclosed in the
narrative when you write it,before you submit it, what your
company name is and who theplayers are.
The next item on what chat GDPor AI devices can't do for you
they can't provide strategy oradvocacy.
Ai can identify discrimination,but it can't file an EOC case.
(23:49):
You need to have a workingknowledge of the EOC process,
which I'll basically give it toyou.
Go to the EOCgov, grab the Form5, fill it out and notarize it.
Say, see attached affidavit,write your fact pattern, get it
notarized and then submit thatthat's your filing of the UOC
charge.
Anything else beyond that it'slike leave the agency to figure
(24:10):
it out.
You know how to deal with it.
But generally they will requestmediation and you want to
submit to that.
They can't negotiate with youremployer.
Well, you need somebody to dothat for you.
An employment lawyer can dothat.
I have to encourage people tonegotiate directly with their
employers and people do it andsuccessfully do it.
It's just simply you know youget your severance package and
(24:32):
they say, no, we're not going torevise that.
They give it to you in a PDFand it's intentional, and you
basically say well, you know,here's my fact pattern, I have
claims here and I'm going to tryto leverage this and push on
you.
You know, make believe to theemployer that you have an
attorney.
You don't have to disclose thename of your attorney to your
employer.
You can just.
You know, if you go throughthis process I've described, you
(24:53):
can come down to a veryprofessionally written thing on
your own.
You may not just use youraffidavit with a Form 5 charge
affixed to it, notarize it.
You can convince your employerbecause the facts in the
affidavit are the most important.
It's not what the lawyer isdoing, it's not what I'm writing
about, it's what the clientsays in the fact pattern.
That's damaging to the employer.
(25:14):
You know, not disclosing.
You have an employment attorneyand just writing a very
cohesive and methodical anddetailed format from A to Z,
chronologically speaking, aboutwhat the claims were.
I've seen clients negotiate formoney and close out a deal
without having to hire anattorney and I love that because
(25:37):
laws, you know it should befree and open.
Why not?
Before there were lawyers,there were just the law.
So lawyers like to curb themarket.
I'm like I'm saying enough ofthe guardrails, let you're
trying to do.
Think smart about it.
Next item the AI can't representyou in court.
So at some jumping off pointyou need to hire an employment
(26:07):
lawyer.
If you want to file a lawsuit,yes, the courts are set up.
The federal courts are set upfor you to file a pro se action.
State courts the same way, andthey will work with you to do
that.
It's a very simplified approach, but you can actually file your
own lawsuit, but you don't wantto avoid that because filing
means notoriety on Googleno-transcript, so be careful.
(26:29):
So at some point you're goingto have to step off and hire an
employment lawyer if you feelyou really need to do that.
So next item is hidden claimsleft on the table.
This one is easy.
Employment lawyers have avariety of claims in addition to
employment claims employmentdiscrimination claims.
I'm sorry that they are lookingat when they're looking at the
(26:51):
same fact pattern.
It's just by experience and soif you don't use an attorney,
you're not going to get thatanalysis to help your situation
and you're going to leave theclaim on the table.
The employer will see itbecause they know about it.
The employers know what they do.
If you don't clue into thataspect, they know exactly what
they're doing.
They want to see if you knowand if you're better at spotting
(27:15):
the issue and seeing whathappened.
That gives you leverage andthat narrative can spell out
that you do know what they did.
That's what the narrative does.
It has to do that.
The other aspect of AI whetherit comes in your own individual
research like this but AIreinforces bias, ai relies on
(27:35):
historical case data.
It may undervalue claims bywomen, minorities or other
groups who already face systemicunderreporting.
The result bad predictionsrooted in past injustices.
Now you know AI.
I don't know if it's smartenough to realize you know
claims that case decisionsoccurred a long time ago.
It's not.
You know whether it's usingcurrent information.
I don't know if it's smartenough to realize claims that
case decisions that occurred along time ago.
It's not whether it's usingcurrent information.
(27:57):
I don't know how they train it,but don't believe the
presumption that they did trainit correctly, because you have
no idea.
I mean they.
I don't know of a productthat's out there yet to let's
provide AI-related efforts tocase decisions to help employees
like yourself.
I don't think that exists.
I'd love to create that toharden your skill set with
(28:20):
knowledge based on cases.
If you want to dive into it,you can, like I said before, you
can research using ChatGDP.
You can go to justiacom andfind reported decisions.
There are other sites thatprovided reported PF decisions
for free of charge.
But the AI reinforces bias issue.
(28:41):
That's the whole argument thathumans created.
The machine learning gave itwhat its source data to absorb,
but there's inherent bias builtinto what it's reading.
The other part of a bias has todo with algorithmic
interviewing.
(29:02):
There's some recent cases we'vetalked about before where the
programming itself, the actualcoding, is biased in some way,
and there's some court decisions, some cases currently ongoing
right now Emotional blind spots.
This issue has to do with youknow if you're going to go all
(29:25):
the way you know, or, to anyextent, with your legal case.
It really is about the issue ofhandling your emotions.
People don't understand thisand they get tripped up really
easily and I have to redirectthem back to center AI.
When you're researching anddeveloping your case, it gives
(29:47):
you cold, technical answers.
I mean it's not saying I'msorry this happened to you.
I've seen other examples whereAI has gotten more emotionally
connected to the user and that'skind of a development I'm
seeing.
I guess the more it knows you,it can get more predictable.
It depends on what device AIdevice you're using.
(30:12):
So the you're trying to navigatethe issue of your employment
discrimination case with youremployer and you get a lot of
fear, anger, anxiety about whatthey're doing to you.
The same goes when you hire meas an employment lawyer.
It's very emotionally charged.
I mean, you name it.
I've seen it.
I've seen great people gothrough the process, gracefully,
(30:34):
delegate to us to do something,manage it, and they understand
and they're managing and they'remanaging expectations.
They're absorbed with theiranxiety about what happened to
them, their own personal issues.
(30:56):
I mean this truism that I knowwhy you got fired.
When I interact with you over aperiod of time or look at your
cases, I can see why youremployer fired you.
If you don't know that, thatyou may be exhibiting behaviors
that tell others and you can'tsee it.
I've seen it.
I mean this is like 28 years.
I've seen exactly what, thebehavior, why people acted the
(31:20):
way they did and they got fired.
So you got to look at theemotionality of what you're
about, to engage, what you'redoing currently, if you're still
working, take a hard look at it.
Don't just assume like you'reperfect and you're normal and
everybody else is effing wrongand they got issues.
So it's something that needs tobe checked.
(31:42):
And I try to draw people intothe transactional analysis of
their situation.
Like, okay, something happenedto you.
Yes, I know it's emotional, butyou're going to dump it all
down in your narrative.
But when you come to thetransaction, using AI to
research and help you researchand present your case.
You want to do it in a verytransactional mindset of, hey,
(32:04):
I'm asking for a year's pay here.
Ok, I'm using this piece ofnarrative that I wrote as
leverage against the employer,that I wrote as leverage against
the employer and I'm doing someresearch here to help me
pinpoint that I'm correct andthey're incorrect about what
happened to me and see if youcan influence them.
Now I can't help this issue.
This one issue is that if youshow up at the table and there's
(32:27):
no reference to who yourattorney is, you might get
called out on it because theemployer is expecting to see
your attorney show up and youhave to fluff this one.
You have to say, well, I'mworking with an attorney and
you're going to go through medirectly.
The employer is going to assumethat you don't have an attorney
and they're going to altertheir dynamics as they're
(32:48):
negotiating with you based uponthat and I know that sucks, but
employers do that.
You have to figure out a way tomake your narrative very, very
convincing, based on actualfacts.
In your research about what'staking place, you may get their
attention and they may want tobury that affidavit that you
(33:09):
provided to them in exchange fora settlement of claims.
So you're very powerful and youhave leverage to work with.
Just slow down.
Take a hard look at what you'rebeing offered in terms of your
chat GDP production results andthink transactionally about this
(33:31):
.
Take the emotions out of it.
If you're wanting moreseverance and you think you have
a claim to do it on, take yourtime and research things.
Maybe there are law clinics andlaw schools.
These folks have students withskills who can assess what
you're doing, possibly help youGhost rate stuff for you.
(33:52):
I've seen this done.
I've seen a lot of stuff done.
Go beyond just what doingpossibly help you Ghostwrite
stuff for you.
I've seen this done.
I've seen a lot of stuff done.
Go beyond just what GDP hasgiven to you.
The AI device is in its infancyand you need to.
Basically, this is about youand your career and what's
happening to you.
So if you're going down thatpipeline, that rabbit hole of
using AI, this whole podcast wasdesigned to make you say wait a
(34:14):
minute, stop, take a breath.
Look what you're using.
Recognize that it's in aninfancy state.
Learn about cases in yourjurisdiction.
That's what lawyers do, youknow.
Supreme Court case decisionsare available for all to see on
the US Supreme Court's website.
Other case decisions come outof various kind of wonky
(34:36):
websites like Justia, but theinformation is out there and
read about these cases and beginto learn about how law is
applied to your set of specificfacts and maybe you might come
out with what the possibilitiesare, you know.
Just to give you a littlefurther cheat, I'm not going to
know the result of every singlecase that I propositioned to an
(34:57):
employer in terms of whatpotentially might happen.
You know the case might justsettle.
I never know.
There's thousands of cases.
I never know the result of whatcould have should have been of
that case.
That's the cheat.
You don't really need to knowthat answer.
What you do need to know iswhat do your facts say?
Pick at AI for a second.
What do your facts say afteryou've spent the time learning
(35:20):
about what sexual harassment isand what hostile work
environment is in the SouthernDistrict of New York under New
York State law and New York Citycode?
By the way, the premier area tobe in both New York City
Southern District, new YorkState, meh, but New York City
definitely.
And if you really do yourresearch, you can be a powerful
(35:40):
device for yourself in yournegotiations with the employer
I've seen it done and forgetabout the future of the case,
what may happen as a result ofwhat a jury will do.
It's not about that.
It's about influencing youremployer to pay you money before
the lawsuit ever gets to thelight of day.
So hopefully that answerssomething for you.
(36:02):
You do have to have a set offacts.
Those facts govern the entiretyof the case.
You know, take your time.
All the information is there.
You've witnessed things and youknow, put it together, make it
compelling, but don't make itthis drama-infused.
You know thing of what youwatch on TV.
It's just matter of fact.
Just say the facts A through Z,chronologically speaking.
(36:25):
Quote people what they said toyou.
You know and go over it andjust you know.
Perfect that fact pattern.
If you're going to do that, useAI to help you, but just have
all these kind of safeguards.
Just you know.
Perfect that fact pattern.
If you're going to do that, useAI to help you, but just have
all these kind of safeguardsthat you know you're going to do
your research and double checkwhat you're doing.
And if eventually you need anemployment lawyer, well, you can
always hire one.
There's lots of them around.
So I hope this helps youunderstand.
(36:48):
You know, using AI from theemployee's perspective some
hidden problems with it, but youstill can use it as a tool to
help you get to where you wantto go.
And really, at the end of theday, it's about severance.
Okay, it's not about gettingyour justice, and people say it
to me.
I'm like I'm sorry.
You know I go about this issuein discussions with judges as
(37:11):
well.
This is the best we can do.
Okay, our justice system.
And it's really about awardingmoney to people through
negotiations and settlement.
Okay, because people, you can'tafford the legal process.
Unfortunately, it's made it toocumbersome financially to do it
.
So there's a way we go throughthis.
(37:31):
Even if you go throughlitigation, we still end up
settling cases before a jurytrial.
Ok, the courts are not.
There's very few jury trials.
You need to understand that.
There's more probably in statecourt, less or so in federal
court.
So you do have access to thesystem in a way I've been
describing.
You do have tools at yourdisposal, which I have gone
through.
(37:51):
You can use AI, but withcaution.
But it's really, at the end ofthe day, it's about you you
writing your facts and youlooking up the law.
Maybe you want to go to lawschool?
Hey, here's a shout out togoing to law school.
Okay, it's rewarding.
You can help other people likeyou in the future if you do it,
and people go to law school allthe time.
But understand, if you want togo to law school for employment
(38:13):
law, you know a worthwhileprofession.
You know there's more need forthem.
Okay, lawyers are aging out,more lawyers are coming in, but
so employment law is veryvaluable to help people like
yourself.
So it's just, you know, maybethat's maybe your calling, maybe
that's what sparked somethinghere for you and you go run with
it.
So, with all that said, that'swhat sparked something here for
you and you go run with it.
(38:34):
So, with all that said, use AIwith caution, know the pitfalls,
but you know, do your research,and it's okay.
You're not going to get all theanswers, but that's okay, and
so hope you found it interesting.
I'll talk to you soon.
Hey, it's Mark and welcome back to another edition
of the Employee Survival Guide.
Today's topic is the hiddendangers of using AI to predict
your employment discriminationcase.
In today's world, employees canturn to artificial intelligence
AI for almost anything, evenlegal advice.
Some employees are beginning touse AI tools to quote-quote
(00:22):
predict whether they have enoughstrong employment
discrimination case.
Employees are tempted with theimmediate feedback that AI
provides in analyzing theirlegal cases.
We have recently seen employeeswho conducted their own legal
research and put together whatthey believe are the correct
legal citations mostly statutesand prominent US Supreme Court
(00:43):
decisions, along with their ownlegal analysis as to why they
have a case of employmentdiscrimination and retaliation.
This is their pitch toemployment lawyers like myself
to review their cases, or theiractual or they're actually
filing it with the state andfederal and city agencies
regarding their claims ofemployment discrimination,
(01:04):
hostile work environment andretaliation.
This came up, this topic came upbecause I receive quite a few
of these type of emails wherepeople are trying to pitch a
story to me and they, you know,no fault of their own.
I mean, ai is out there.
It's new and novel andemployers are using it, of
(01:28):
course, but you know, peoplehave access to a Google notebook
or a chat, gdp or something,and the technology itself is
relatively new and it's basedupon you know, upon what you fed
into the system.
And so when I read these emailsand they give me the cut and
(01:50):
paste of what their research wasand we all know what this looks
like it's like this chat orthis AI type of production of
work product that it looks likeit's usually a list or that's
kind of stale, it's not veryemotional, it's just a matter,
or that's kind of stale, it'snot very emotional, it's just
matter of fact, whatever.
But it's only based upon whatthe AI device has captured in
(02:11):
its learning.
And the funny thing about thistopic is that the AI devices,
these algorithms, can pick upinformation related to agency
websites and devour that, butthe case law material, the case
decisions from our courts statecourts, federal courts they're
(02:33):
not readily accessible andthey're behind paywalls where
you have to go in and access it.
Now there are cases reportedout there that there are various
websites that put it out, butthe vast majority of case
decisions are locked behindpaywalls for websites called
like Westlaw and LexisNexis, andso the AI devices that the
(02:57):
employees are accessing don'thave that wealth of depth so
that information is missing fromits analysis.
Now, this is a technology ininfancy, so you can imagine that
if employees are puttingtogether their case analysis
today, it might get more robustas the technology increases,
(03:18):
maybe what they feed into it.
I would love to see the AIdevices like a chat GDP, have
all of the case materialdecisions going back, let's say,
just give it 10 years.
I mean these devices can learnthe machine learned very, very
quickly device so the employeescould access it, because the
(03:45):
employees would basically beable to read public court
decisions because courts arepublic and read these decisions
and decide for themselves ifthey have a case or not.
Remember, lawyers are not theonly parties that have an angle
or the access to the law.
Everybody has access to the lawand at least that's my opinion
(04:07):
and that's why I try to put theslant I do on these podcast
episodes.
I give you case decisions.
You can hear about it.
So the so having employeesaccess in case decisions is
really uber important toleveling the playing field
(04:29):
between employers and employees,which is not even, which is
uneven, the whole at-will debateyou hear about me raise all the
time.
So back to the point ofemployees trying to figure out
their cases before they confirma lawyer or file before an
agency and they're doing theirAI research.
I'm not saying don't do this.
I'm saying just do this withthe eyes wide open about what
(04:51):
you're looking at, and it's okayto begin to learn like a lawyer
, and this podcast is reallydesigned to kind of tell you
some pitfalls, but also tell youhow do you approach it to do it
better, but also tell you youknow how do you approach it to
do it better.
You know how do you take yourAI research and create it in
such a way that it's moreaccurate than it wasn't.
(05:13):
So first is you know, read Igot a lot of material on my
website and a lot of podcaststalk about it.
For example, hostile at WorkEnvironment Everybody loves that
topic on my website and thepodcast, and so there's a lot of
information about it.
(05:33):
For example, hostile workenvironment Everybody loves that
topic on my website and in thepodcast, and so there's a lot of
information about that.
But how do you interpret ahostile work environment with
respect to the you know, a factpattern that you, what you're
currently going through.
So what I encourage you to dois this why don't you write out
your fact pattern and put it inlong form you know your
narrative from A to Z,chronological order, what
happened, what happened next,who knew about it, whatever, and
write it out?
(05:54):
You might want to not includenames of the employer and use
fictional names of characters,because you're putting this
information publicly out thereinto a domain AI device that's
learning it and you haveconfidentiality provisions of
your when you sign up foremployment, so you would be
careful what you're putting outthere.
(06:14):
I mean, it's like the blackhole, this AI device.
So just caution.
You Just write a narrative offacts and then put it into the
algorithm, the chat GDP, and askit to assess based upon, let's
say, you live in New York Cityand you say AI chat GDP, why
(06:34):
don't you assess it under NewYork City law, human rights law,
new York State human rights lawand Title VII, 1964 Civil
Rights Act in New York State,and let's see how that chat GDP
device pushes out the outcome.
Now we can't trust what it'sgoing to say, but hopefully, if
(06:57):
you're using your fact pattern,you might get a closer vantage
point to illegality, toillegality.
Sometimes when I do the searcheson chat GDP.
I'll get warnings that they'renot a lawyer and they can't make
legal or give legal advice.
You might see that as well, butplay around with it and you
might be able to come very closeto, let's say, your fact
(07:18):
pattern, directing you to theconclusion that you have a
hostile work environment.
And everybody knows what ahostile work environment is.
It's, you know, a pervasive useof discriminatory comments or,
you know, comparison of youversus others that are given
you've received more favorabletreatment.
It's typically in a sexualconnotation or sometimes it's in
(07:41):
a racial connotation where youhave, like you know, pictures of
nooses for a racial case orlightning bolts for the KKK or
some pervasive hostile likeextreme situation which is, and
maybe the computer can spit outthat assessment for you.
(08:01):
So go through that exercise,begin to learn more.
If you spend the time, you canfigure out that there's
information out there and I'mgoing to tell you this.
The next chat GDP search shouldbe give me notable let's say
Southern District of New Yorkcases in the last five years
(08:23):
involving hostile workenvironment related to sex and
see what those decisions come upwith.
And let's say you get five, sixdecisions and you push those
case names back into a Googlesearch and see if you can come
across like a Cornell decision.
They have a website that hascase decisions.
Justia is another site that has, you know, pdfs.
(08:46):
You can read the actualdecision, but use the devices,
the search, in a way that youcan use ChatGP to identify cases
.
You can use a variety ofsources to identify cases on
sexual or hostile environment inthe Southern District of New
York.
It's a federal court and see ifyou can find the cases
themselves to read through themand see if you can find the
(09:07):
cases themselves to read throughthem.
What you're trying to do here isread through the narrative of
fact in the decision, but thenalso look at the court's
analysis of that, and that'swhere I really want to bring
your attention.
What's wrong with the AI use byemployees today is that the
employees don't have anyunderstanding of the analysis as
(09:27):
it relates to being a lawyerand giving the assessment.
Now, I'm not saying thatemployees have to become quick,
off-the-cuff attorneys toidentify their cases.
It's much simpler than that.
I think you have the ability tolook at here's what a case
decision is from a judge andsays here's the law I'm applying
to this narrative of this casedecision you just found using
(09:50):
research, and you can see howthe judge moves through that
analysis, using the fact andthen applying the law to the
fact to establish the conclusion.
Well, there you have it.
You've just gone to law school.
You've learned what the factsare, you've learned the law part
of it and you've learned theanalysis fact.
In law that's called issuespotting or I don't want to give
(10:13):
anything to law school.
It's just looking at facts,looking at law, seeing how the
facts apply and how the law mayapply, and it's a very generic
way of looking at the strength,weakness of a case.
So it's a very generic way oflooking at the strength,
weakness of a case.
Now you would have to see a lotmore of those cases to decide
well, do I have a case or not,or is it how strong it is?
I think you're better offwriting a fact pattern and using
(10:37):
your research, saying, ok, Ifound a sexual harassment case
with a hostile work analysis byJudge so-and-so, and I put that
into my analysis and maybe I'mgoing to search further because
I want to find some more cases.
And here's a trick Once I startto find a commonality of case
decisions about the same thingin different court cases in the
(11:01):
same district, let's say theSouthern District of New York,
which I'm admitted to.
I stop and I realize I have thebasic foundation of what sexual
harassment, hostile workenvironment is for the Southern
District and you want to movefurther into the analysis and
you look at possible otherclaims that might exist and use
(11:21):
the case decision you're readingor researching to identify more
cases.
Now I'll give you a cheat okay,employment discrimination cases
fall into the protectedcategory.
Okay, and it's age, sex,whatever race, religion,
(11:41):
national origin, disability, andyou know, if you're, by and
large, you're going to fall intoone of those buckets or
retaliation or whistleblowing,and you want to do that search
by those different categories ofdiscrimination to find cases
that are applicable to yours.
Now, you're not going to findthe exact case applicable on
point.
They call it.
That takes a little bit deeperdive into the case law, spending
hours trying to locate cases,but you're going to get a quick
(12:03):
assessment using AI, usingGoogle search, using Justia, the
website, to locate cases toread.
You have to read this stuff.
Don't expect ChatGDP to push outthe analysis for you like we're
used to, because that's not theway it works.
What are you missing?
You're missing the subtletiesthat the law is.
You're missing the subtletiesto which an employment lawyer
(12:25):
will give you.
But I think you can come close.
I think you can educateyourself about the laws, because
I've seen clients do thisbefore AI got involved.
I've seen clients go and do theresearch at law schools and law
libraries and they came prettyclose to what the actual
analysis is.
So there's a lot of commonalityof case claims that you can pick
(12:50):
up very quickly to put in.
I know that ChatGDP will pushout.
Here's what sexual or hostilework environment will mean, and
you'll start to.
You know, because you put it inseveral times, you get the same
answer.
That's not the case.
If you actually read a courtdecision.
By doing your deep diveresearch, you can actually
figure out how the courts arelaying out the elements of the
claim for hostile workenvironment.
(13:13):
Then the next stage of that iswell, step back and look at your
fact pattern in the analysisand see, you know, are there
inferences in your fact patternthat support the possibility
that you have a claim of hostileworkment discrimination?
We're looking for stronginferences.
So we're going to take all.
(13:34):
There's two types of evidencehere.
There's direct statements,statements we don't like you
because you know you're an f andwhatever, and uh, and they
start to do extreme things inthe office and there's uh, just
hostility, whatever, uh, andthen there's that's all direct
statement, stuff from a boss, uh.
And then you have this andthat's like five percent of the
cases.
So it's out there but it'ssmoking on the evidence and it's
(13:56):
not really what you'retargeting.
You're targeting circumstantialevidence.
It's how you're treateddifferently than other people
who are treated more favorably,who are the different categories
.
So, if it's sex, hostile workenvironment you're looking at,
and you're female, you'relooking at someone who's a male,
who's treated better, and youwant to do this compare and
contrast, assessment in your ownfact pattern, looking for
(14:19):
inferences of how they'retreated differently.
Assessment in your own factpattern, looking for inferences
of how they're treateddifferently.
I mean, that's really where thepart of I want you to pick up
on.
You know, look at chat GDP, getyour data, your information
about what the law is, do someadditional research to confirm
you're correct, because AI isnot perfect yet.
Get an understanding of workingknowledge of what hostile work
environment is, look at yourfact pattern and then begin to
(14:42):
do your analysis, I think fivethings stand out.
Five circumstantial pieces ofevidence point out of compare
and contrast, and I was firedbecause I didn't receive the
same treatment they did.
Let's not make it complicated.
Okay, put your facts together.
Do your research.
What's going to happen to youis when you do your research and
you spend the time reading,like lawyers do in these cases,
(15:04):
or listening to a podcast abouthostile work environment on my
podcast, you're going to beginto become familiar, like with
anything, the subject matterthat you're working with, and
you'll start to see the nuancesof things.
You'll start to see things popup for you factually that you
may want to put back into yournarrative.
Spend some time.
Don't just shotgun the approachof saying here's a stretch GDP,
(15:26):
I'm going to put my informationand, bang, I'm going to send
that to the lawyer.
Okay, I'm not kidding, thatdoes happen.
It irritates me.
What I'm trying to say is youcan figure out and come pretty
darn close to what your case isif you go about this process of
assessing what lawyers do.
They ask people to write thefact pattern, write the
(15:47):
narrative chronologically.
Don't provide conclusions, juststate the facts, ma'am, and
then do some research andunderstanding what this is?
Because if you're going to aska lawyer to take a case on
contingency which is even worseask a lawyer.
If you're going to ask a lawyerto take a case on contingency
which is even worse you want tobe convincing.
And if you're not convincing inyour initial you know email,
(16:12):
it's going to take about twoseconds for me to realize it and
I'm going to basically dispensewith the.
You know I'm not going to writeback to you, I'm just going to
say that's not a case interestspending my time on.
If you're going to hire a lawyeron an hourly basis and you're
throwing your own money at it,you want to know that you're
hiring an employment lawyer andyou're spending your money
wisely.
There you become acollaborative effort with the
(16:34):
employment attorney and I'veseen clients who are bright and
they're pretty savvy and theycan.
They work in a kind of acollaborative effort to push
along a case and they understandit and so you can age yourself
in the process of understandingthe legal services you're
receiving and understand yourcase.
(16:55):
But if you go about this aspectof the strategy of pushing
things into chat GDP, it's goingto be met with like false left
and right red flags.
People pick it up prettyquickly.
Employers and their counsel canread it very quickly, say oh,
that was obviously AI produced,et cetera.
So people are doing that.
(17:15):
But here's the things to avoid.
To get down to the nitty grittyand it's you know, when you're
creating your narrative andyou're doing your research and
you're going the extra mile tofigure out what it is and it is
not sexual harassment, hostileenvironment.
Beware of the following issues.
Okay, ai oversimplifies the law.
Implement discrimination iscomplex, fact-specific and
(17:36):
constantly evolving.
Complex, fact-specific andconstantly evolving.
Now I told you kind of thebasic primer of how to look at
these cases in terms of yourclaims, because I'm doing that
deliberately, I want to dumb itdown, because I want to make it
more accessible to you.
The law has a tendency to makeit really complex, but I'm
trying to make it easier toabsorb and there are a lot of
(18:01):
commonalities amongst claims andyou can only see that when you
look at it long enoughexperiencing it.
So AI tends to be applied broad,rigid rules, missing exceptions
or subtleties that I wouldunderstand as an employment
lawyer, and they are exceptionsthat might make or break your
case.
I'll give you an example.
(18:24):
So you have a claim.
You are 365 days, 366 days,since you were fired, and it's
well, and you decide finally todo something about it.
What you don't know is thatwhen you filed your UOC charge,
you were a little bit latebecause the statute of
limitations says 300 days, soyou wouldn't know is that when
you filed your USC charge, youwere a little bit late because
the statute of limitations says300 days, so you wouldn't know
(18:45):
that.
So that's an exception.
So just AI oversimplifies thelaw.
So be very, very careful.
Next thing is wrong predictions, wrong decisions.
I start this off with the famousset of cases that came out of
the Southern District of NewYork.
This involved lawyers.
(19:05):
It's a true story.
It was reported in the pressTwo cases of lawyers who were
before the federal judges in twoseparate cases.
Both had conducted AI researchon legal cases, had conducted AI
research on legal cases and thejudge got a little bit upset in
(19:27):
both cases and said and theywere sanctioned because the AI
that they were using whateverthat was had made up case names
and made up case decisionalanalysis, meaning made it up.
I don't know what they wereusing, but both lawyers in both
cases got reprimanded.
And it's very embarrassing.
The judges look at every singlecitation that we give to them
(19:49):
because we're looking at thecases and saying this is
supportive of our client's case.
So, likewise, I'm asking you todo the same thing when you do a
research through ChatGDP Followup and read the case.
You know these lawyers in thatcase that I'm citing the two
cases they did not read thedecisions that this research
chatbot had created for them.
(20:09):
It's embarrassing, but itproves the point that even
lawyers could do the samescrew-up mistake.
So don't do that.
So that's an example ofnon-existing cases and
non-existing case decisionanalysis.
So AI isn't trained on yourunique facts or the latest court
(20:30):
rulings.
That means it can give falseconfidence quote I will win the
case or unnecessarydiscouragement.
I don't have a case leading youto make damaging choices.
So AI is in its infancy.
Like I explained before, itdoesn't have all the thousands
and hundreds of thousands ofcase decisions across the
(20:51):
country in its data bank becauseit's beyond a paywall.
Yeah, you're like.
Well, that's unfair.
Why is that?
Courts are open.
Why can't we access it?
Well, it's a larger question.
I think that we should haveaccess to all the case decisions
in both state and federalcourts.
But maybe there's a reason whythat's not the case, right,
(21:11):
because more information ispower the more power you have,
the more damage you can do on acorporation, so maybe that's the
case.
There's also Westlaw and Lexiswho serve this services, and I
pay a good dollar for thatservice my lawyers myself.
The next item for what AI is notto be trusted is missed
(21:31):
deadlines and lost rights, likethe example I gave you.
If you come up with your hey, Iwant to file a new EOC case 365
days after the adverse actionmeaning you got fired you're too
late.
365 days after the adverseaction meaning you got fired,
you're too late.
You got to file your decisionor your claim to exhaust your
administrative remedies in 300days or within 300 days, so
easiest is missed claims becauseyou had.
(21:54):
You didn't understand thefiling requirements in terms of
deadlines.
I'll give you a basic primer180 days and 300 days on
employment discrimination cases.
Okay, let's just follow that.
So the adverse actions have tooccur within that.
Most state and federal agencieswork on what's called a
work-sharing agreement and soyou get the benefit of a 300-day
window, and I would alwaysencourage you to file with the
(22:17):
EOC first and then have the EOCfile the claim with the local
state agency.
So you get the 300-day windowtime period.
The next item I spoke aboutbefore is when you create your
written narrative and you put itinto the chat GDP to figure it
out.
Use generic names.
(22:40):
Don't use company informationthat's identifying of the
company.
When you type your story into apublic AI system, you expose
sensitive details names, dates,medical history that you can't
take it back.
Ai doesn't guarantee a privacyand you sign most likely
confidentiality agreements withyour employers not to share the
(23:04):
information publicly.
You're putting the informationin a chat GDP or an AI device.
You have disseminatedinformation out there publicly.
So be very careful because it'snot about if.
It's about when these systemsand employers catch up that they
can figure out who disclosedthe information.
So be generic with your namesand company information and what
(23:26):
you're talking about, so itcan't be disclosed in the
narrative when you write it,before you submit it, what your
company name is and who theplayers are.
The next item on what chat GDPor AI devices can't do for you
they can't provide strategy oradvocacy.
Ai can identify discrimination,but it can't file an EOC case.
(23:49):
You need to have a workingknowledge of the EOC process,
which I'll basically give it toyou.
Go to the EOCgov, grab the Form5, fill it out and notarize it.
Say, see attached affidavit,write your fact pattern, get it
notarized and then submit thatthat's your filing of the UOC
charge.
Anything else beyond that it'slike leave the agency to figure
(24:10):
it out.
You know how to deal with it.
But generally they will requestmediation and you want to
submit to that.
They can't negotiate with youremployer.
Well, you need somebody to dothat for you.
An employment lawyer can dothat.
I have to encourage people tonegotiate directly with their
employers and people do it andsuccessfully do it.
It's just simply you know youget your severance package and
(24:32):
they say, no, we're not going torevise that.
They give it to you in a PDFand it's intentional, and you
basically say well, you know,here's my fact pattern, I have
claims here and I'm going to tryto leverage this and push on
you.
You know, make believe to theemployer that you have an
attorney.
You don't have to disclose thename of your attorney to your
employer.
You can just.
You know, if you go throughthis process I've described, you
(24:53):
can come down to a veryprofessionally written thing on
your own.
You may not just use youraffidavit with a Form 5 charge
affixed to it, notarize it.
You can convince your employerbecause the facts in the
affidavit are the most important.
It's not what the lawyer isdoing, it's not what I'm writing
about, it's what the clientsays in the fact pattern.
That's damaging to the employer.
(25:14):
You know, not disclosing.
You have an employment attorneyand just writing a very
cohesive and methodical anddetailed format from A to Z,
chronologically speaking, aboutwhat the claims were.
I've seen clients negotiate formoney and close out a deal
without having to hire anattorney and I love that because
(25:37):
laws, you know it should befree and open.
Why not?
Before there were lawyers,there were just the law.
So lawyers like to curb themarket.
I'm like I'm saying enough ofthe guardrails, let you're
trying to do.
Think smart about it.
Next item the AI can't representyou in court.
So at some jumping off pointyou need to hire an employment
(26:07):
lawyer.
If you want to file a lawsuit,yes, the courts are set up.
The federal courts are set upfor you to file a pro se action.
State courts the same way, andthey will work with you to do
that.
It's a very simplified approach, but you can actually file your
own lawsuit, but you don't wantto avoid that because filing
means notoriety on Googleno-transcript, so be careful.
(26:29):
So at some point you're goingto have to step off and hire an
employment lawyer if you feelyou really need to do that.
So next item is hidden claimsleft on the table.
This one is easy.
Employment lawyers have avariety of claims in addition to
employment claims employmentdiscrimination claims.
I'm sorry that they are lookingat when they're looking at the
(26:51):
same fact pattern.
It's just by experience and soif you don't use an attorney,
you're not going to get thatanalysis to help your situation
and you're going to leave theclaim on the table.
The employer will see itbecause they know about it.
The employers know what they do.
If you don't clue into thataspect, they know exactly what
they're doing.
They want to see if you knowand if you're better at spotting
(27:15):
the issue and seeing whathappened.
That gives you leverage andthat narrative can spell out
that you do know what they did.
That's what the narrative does.
It has to do that.
The other aspect of AI whetherit comes in your own individual
research like this but AIreinforces bias, ai relies on
(27:35):
historical case data.
It may undervalue claims bywomen, minorities or other
groups who already face systemicunderreporting.
The result bad predictionsrooted in past injustices.
Now you know AI.
I don't know if it's smartenough to realize you know
claims that case decisionsoccurred a long time ago.
It's not.
You know whether it's usingcurrent information.
I don't know if it's smartenough to realize claims that
case decisions that occurred along time ago.
It's not whether it's usingcurrent information.
(27:57):
I don't know how they train it,but don't believe the
presumption that they did trainit correctly, because you have
no idea.
I mean they.
I don't know of a productthat's out there yet to let's
provide AI-related efforts tocase decisions to help employees
like yourself.
I don't think that exists.
I'd love to create that toharden your skill set with
(28:20):
knowledge based on cases.
If you want to dive into it,you can, like I said before, you
can research using ChatGDP.
You can go to justiacom andfind reported decisions.
There are other sites thatprovided reported PF decisions
for free of charge.
But the AI reinforces bias issue.
(28:41):
That's the whole argument thathumans created.
The machine learning gave itwhat its source data to absorb,
but there's inherent bias builtinto what it's reading.
The other part of a bias has todo with algorithmic
interviewing.
(29:02):
There's some recent cases we'vetalked about before where the
programming itself, the actualcoding, is biased in some way,
and there's some court decisions, some cases currently ongoing
right now Emotional blind spots.
This issue has to do with youknow if you're going to go all
(29:25):
the way you know, or, to anyextent, with your legal case.
It really is about the issue ofhandling your emotions.
People don't understand thisand they get tripped up really
easily and I have to redirectthem back to center AI.
When you're researching anddeveloping your case, it gives
(29:47):
you cold, technical answers.
I mean it's not saying I'msorry this happened to you.
I've seen other examples whereAI has gotten more emotionally
connected to the user and that'skind of a development I'm
seeing.
I guess the more it knows you,it can get more predictable.
It depends on what device AIdevice you're using.
(30:12):
So the you're trying to navigatethe issue of your employment
discrimination case with youremployer and you get a lot of
fear, anger, anxiety about whatthey're doing to you.
The same goes when you hire meas an employment lawyer.
It's very emotionally charged.
I mean, you name it.
I've seen it.
I've seen great people gothrough the process, gracefully,
(30:34):
delegate to us to do something,manage it, and they understand
and they're managing and they'remanaging expectations.
They're absorbed with theiranxiety about what happened to
them, their own personal issues.
(30:56):
I mean this truism that I knowwhy you got fired.
When I interact with you over aperiod of time or look at your
cases, I can see why youremployer fired you.
If you don't know that, thatyou may be exhibiting behaviors
that tell others and you can'tsee it.
I've seen it.
I mean this is like 28 years.
I've seen exactly what, thebehavior, why people acted the
(31:20):
way they did and they got fired.
So you got to look at theemotionality of what you're
about, to engage, what you'redoing currently, if you're still
working, take a hard look at it.
Don't just assume like you'reperfect and you're normal and
everybody else is effing wrongand they got issues.
So it's something that needs tobe checked.
(31:42):
And I try to draw people intothe transactional analysis of
their situation.
Like, okay, something happenedto you.
Yes, I know it's emotional, butyou're going to dump it all
down in your narrative.
But when you come to thetransaction, using AI to
research and help you researchand present your case.
You want to do it in a verytransactional mindset of, hey,
(32:04):
I'm asking for a year's pay here.
Ok, I'm using this piece ofnarrative that I wrote as
leverage against the employer,that I wrote as leverage against
the employer and I'm doing someresearch here to help me
pinpoint that I'm correct andthey're incorrect about what
happened to me and see if youcan influence them.
Now I can't help this issue.
This one issue is that if youshow up at the table and there's
(32:27):
no reference to who yourattorney is, you might get
called out on it because theemployer is expecting to see
your attorney show up and youhave to fluff this one.
You have to say, well, I'mworking with an attorney and
you're going to go through medirectly.
The employer is going to assumethat you don't have an attorney
and they're going to altertheir dynamics as they're
(32:48):
negotiating with you based uponthat and I know that sucks, but
employers do that.
You have to figure out a way tomake your narrative very, very
convincing, based on actualfacts.
In your research about what'staking place, you may get their
attention and they may want tobury that affidavit that you
(33:09):
provided to them in exchange fora settlement of claims.
So you're very powerful and youhave leverage to work with.
Just slow down.
Take a hard look at what you'rebeing offered in terms of your
chat GDP production results andthink transactionally about this
(33:31):
.
Take the emotions out of it.
If you're wanting moreseverance and you think you have
a claim to do it on, take yourtime and research things.
Maybe there are law clinics andlaw schools.
These folks have students withskills who can assess what
you're doing, possibly help youGhost rate stuff for you.
(33:52):
I've seen this done.
I've seen a lot of stuff done.
Go beyond just what doingpossibly help you Ghostwrite
stuff for you.
I've seen this done.
I've seen a lot of stuff done.
Go beyond just what GDP hasgiven to you.
The AI device is in its infancyand you need to.
Basically, this is about youand your career and what's
happening to you.
So if you're going down thatpipeline, that rabbit hole of
using AI, this whole podcast wasdesigned to make you say wait a
(34:14):
minute, stop, take a breath.
Look what you're using.
Recognize that it's in aninfancy state.
Learn about cases in yourjurisdiction.
That's what lawyers do, youknow.
Supreme Court case decisionsare available for all to see on
the US Supreme Court's website.
Other case decisions come outof various kind of wonky
(34:36):
websites like Justia, but theinformation is out there and
read about these cases and beginto learn about how law is
applied to your set of specificfacts and maybe you might come
out with what the possibilitiesare, you know.
Just to give you a littlefurther cheat, I'm not going to
know the result of every singlecase that I propositioned to an
(34:57):
employer in terms of whatpotentially might happen.
You know the case might justsettle.
I never know.
There's thousands of cases.
I never know the result of whatcould have should have been of
that case.
That's the cheat.
You don't really need to knowthat answer.
What you do need to know iswhat do your facts say?
Pick at AI for a second.
What do your facts say afteryou've spent the time learning
(35:20):
about what sexual harassment isand what hostile work
environment is in the SouthernDistrict of New York under New
York State law and New York Citycode?
By the way, the premier area tobe in both New York City
Southern District, new YorkState, meh, but New York City
definitely.
And if you really do yourresearch, you can be a powerful
(35:40):
device for yourself in yournegotiations with the employer
I've seen it done and forgetabout the future of the case,
what may happen as a result ofwhat a jury will do.
It's not about that.
It's about influencing youremployer to pay you money before
the lawsuit ever gets to thelight of day.
So hopefully that answerssomething for you.
(36:02):
You do have to have a set offacts.
Those facts govern the entiretyof the case.
You know, take your time.
All the information is there.
You've witnessed things and youknow, put it together, make it
compelling, but don't make itthis drama-infused.
You know thing of what youwatch on TV.
It's just matter of fact.
Just say the facts A through Z,chronologically speaking.
(36:25):
Quote people what they said toyou.
You know and go over it andjust you know.
Perfect that fact pattern.
If you're going to do that, useAI to help you, but just have
all these kind of safeguards.
Just you know.
Perfect that fact pattern.
If you're going to do that, useAI to help you, but just have
all these kind of safeguardsthat you know you're going to do
your research and double checkwhat you're doing.
And if eventually you need anemployment lawyer, well, you can
always hire one.
There's lots of them around.
So I hope this helps youunderstand.
(36:48):
You know, using AI from theemployee's perspective some
hidden problems with it, but youstill can use it as a tool to
help you get to where you wantto go.
And really, at the end of theday, it's about severance.
Okay, it's not about gettingyour justice, and people say it
to me.
I'm like I'm sorry.
You know I go about this issuein discussions with judges as
(37:11):
well.
This is the best we can do.
Okay, our justice system.
And it's really about awardingmoney to people through
negotiations and settlement.
Okay, because people, you can'tafford the legal process.
Unfortunately, it's made it toocumbersome financially to do it
.
So there's a way we go throughthis.
(37:31):
Even if you go throughlitigation, we still end up
settling cases before a jurytrial.
Ok, the courts are not.
There's very few jury trials.
You need to understand that.
There's more probably in statecourt, less or so in federal
court.
So you do have access to thesystem in a way I've been
describing.
You do have tools at yourdisposal, which I have gone
through.
(37:51):
You can use AI, but withcaution.
But it's really, at the end ofthe day, it's about you you
writing your facts and youlooking up the law.
Maybe you want to go to lawschool?
Hey, here's a shout out togoing to law school.
Okay, it's rewarding.
You can help other people likeyou in the future if you do it,
and people go to law school allthe time.
But understand, if you want togo to law school for employment
(38:13):
law, you know a worthwhileprofession.
You know there's more need forthem.
Okay, lawyers are aging out,more lawyers are coming in, but
so employment law is veryvaluable to help people like
yourself.
So it's just, you know, maybethat's maybe your calling, maybe
that's what sparked somethinghere for you and you go run with
it.
So, with all that said, that'swhat sparked something here for
you and you go run with it.
(38:34):
So, with all that said, use AIwith caution, know the pitfalls,
but you know, do your research,and it's okay.
You're not going to get all theanswers, but that's okay, and
so hope you found it interesting.
I'll talk to you soon.
Popular Podcasts
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.
Cardiac Cowboys
The heart was always off-limits to surgeons. Cutting into it spelled instant death for the patient. That is, until a ragtag group of doctors scattered across the Midwest and Texas decided to throw out the rule book. Working in makeshift laboratories and home garages, using medical devices made from scavenged machine parts and beer tubes, these men and women invented the field of open heart surgery. Odds are, someone you know is alive because of them. So why has history left them behind? Presented by Chris Pine, CARDIAC COWBOYS tells the gripping true story behind the birth of heart surgery, and the young, Greatest Generation doctors who made it happen. For years, they competed and feuded, racing to be the first, the best, and the most prolific. Some appeared on the cover of Time Magazine, operated on kings and advised presidents. Others ended up disgraced, penniless, and convicted of felonies. Together, they ignited a revolution in medicine, and changed the world.