More and more job applications are processed by machine learning before a real person ever reads them. But can these algorithms exhibit prejudice? And, if so, what would it mean to adopt algorithmic affirmative action?
In this episode, we sit down with Stetson Professor Jason Bent to discuss the changing landscape of employment and employment discrimination law in the twenty-first century. We discuss the impact of AI, growing concerns about data privacy in employment contexts, and the role new Supreme Court decisions have played in the interpretation of Title VII.
Transcript:
Speaker 1 (00:02.808)
There were three different opinions and all three of them claimed to be faithfully applying the ordinary meaning of the words because of sex in Title VII, which can't possibly be true, right? Someone needs to be right about this, but they all were taking the textualist approach. They all said the text compels this result. Justice Gorsuch wrote the majority opinion. says it necessarily, includes logically, it includes sexual orientation, even if people didn't think
that it included sexual orientation. The words they chose actually did include sexual orientation. And Justice Kavanaugh writes a dissent and he says he's taking the textualist approach as well. But he says in the ordinary parlance, people thought of sex as something different than sexual orientation. And so they included sex as a thing that you can't discriminate on, but they didn't include sexual orientation. So they made a choice between those.
to essentially. So everyone claims to be a textualist but they reach totally different results.
This is Real Cases, a legal podcast presented by the Stetson University College of Law. We'll sit down with Stetson Law faculty and students to examine today's critical cases and debates in environmental, international, elder, and business law, plus the role of social justice in these fields. Join us as we open the case file.
Episode 30, Employment Law and Discrimination. I'm Daniel O'Keefe, Master of English Literature from Indiana University. Today we're joined by Professor Jason Bent. Professor Bent focuses on employment law, employment discrimination law, and civil procedure. His scholarly interests include systemic theories of employment discrimination, federal workplace safety regulation, and the use of economic theory and statistical techniques in the development of legal doctrine.
Speaker 2 (02:01.528)
He is the co-author of An Illustrated Guide to Civil Procedure and The Statistics of Discrimination, Using Statistical Evidence in Discrimination Cases. His work has been featured in a range of publications, including the Georgetown Law Journal, the Ohio State Law Journal, and the Brigham Young University Law Review. So I'd like to start out by asking first just what initially got you interested in employment law? How did you end up focusing on that yourself?
Great question. I like to tell my students a sort of humorous little story about that. I didn't take employment discrimination. I didn't take employment law. take labor law. Didn't take employee benefits. Didn't take any of those courses in law school. Had no inkling that that would be an area I might pursue. But then I did a judicial clerkship in the Northern District of Illinois. So that was in Chicago for a district judge there named Judge Gottschall, Judge Joan Gottschall.
And a decent amount of the, the decent portion of the docket in the federal courts generally, but in the Northern District of Illinois, maybe in particular, is employment discrimination suits. And so I got to work on several different really interesting employment discrimination suits when I was a clerk. I hadn't really been exposed to that area of law much before that, a little bit with some assignments at the firm, but not much before that.
But I found that I really loved those cases and my co-clerk and I sort of worked out how we divided up cases according to subject matter. So he would take all the patent cases and I would take all the employment discrimination cases. And then when I, after my clerkships, I did two clerkships and then after my clerkships, when I joined the firm, I asked to be put on employment discrimination cases and then the rest is history. I sort of ended up in the employment practice there.
interesting. So I feel like this might be a very naive sort of question, but I'm curious nonetheless. So.
Speaker 2 (04:06.85)
The majority of employment discrimination suits, how do they arise? And I ask this in light of this particular issue. I feel like discrimination so often is something where people have an inkling that it may have occurred, but they have no real firm thing to put, to say with certainty, I feel absolutely confident that I was discriminated against and I can demonstrate it this way. And so I feel like
it when it comes to at least down to individuals in any particular case of not getting a job or being fired, I feel like very often people must be in a very hazy sort of situation when it comes to determining whether or not they were actually discriminated against. So I'm curious about how these suits arise and how they usually unfold.
I mean, I think you're exactly right about that. Circumstantial evidence is really what most employment discrimination cases boil down to because employers understand the law. They don't usually provide direct evidence that they're discriminating against someone on the basis of a protected category. I can't say they never do, but they usually don't. And so what happens is an employee
suffers an adverse action, oftentimes they get fired from a job. And then, you know, I think they reflect at that time on the things that they can, that they have in their knowledge that they can point to that suggest maybe the decision maker was biased. And so those can be things like,
comments that were made to them by coworkers or supervisors, stories that were told to them by a coworker about a supervisor, or it can be that they observed a pattern over time. Then they can point to the pattern. But ultimately, a lot of times they can't get to the best evidence of discrimination until they actually get into
Speaker 1 (06:20.832)
either an EEOC investigation where they might learn some things or the lawsuit. Once they file a civil suit, they'll have, if they can get to discovery, get past the motion to dismiss and get into discovery, then they'll get a chance to actually depose the decision-makers, get to see some files. They will, in discovery, they will a lot of times try to push for evidence that might suggest a pattern. So statistics over time of what happened to other individuals.
Speaker 1 (06:53.609)
That's one way that these cases arise is there are in a bit of a hazy situation, but they have some inkling based on something that was said to them or some pattern that they saw. Then they may sue. Now, another thing that can happen and that is a trend in employment discrimination is maybe they were harassed or a custom comment was made at work, then they-
reported it, they complained about it to the employer and to the employer, whether the HR or otherwise. So they've made some report. They've engaged in what the statute calls protected activity. Well, what we call, but employment discrimination practitioners would call protected activity. And the statute says, if you oppose or unlawful action, then that is protected and the employer can't retaliate against you because of it. So maybe someone
feels that they've been suffering some kind of harassment at work or that they've observed someone else suffering some kind of harassment at work, then they make a complaint about it. And then they suffer an adverse action, like not getting a promotion, getting demoted, getting fired. And they have a retaliation claim. And actually retaliation claims are the most commonly charged kinds of discrimination. it's relatively easy to, you get fired after you made a complaint of race discrimination, for example, or racial harassment.
Alright.
Speaker 1 (08:16.344)
A lot of times the employee will say it was both retaliation, but then also racially discriminatory. I was fired because of my race and also I was fired because I complained about harassment based on my race.
Got it, okay. Now, both of, so you said that retaliation complaints are the most common.
They're the most common charge at this point, yeah.
And both of the instances that you brought up just now had to do specifically with someone who is employed and then suffers some sort of adverse situation. How about people who are trying to make the case that they weren't hired in the first place on a calendar race?
Yeah.
Speaker 1 (08:56.398)
extraordinarily hard to make that case because it's basically a black box to you. You applied, maybe you got an interview, but you didn't get hired. You don't know why. Maybe they gave you some, there was another more qualified candidate or something, but you don't have any real anything to point to to show that it might have been sex or race discrimination or religious discrimination unless something happened during the interview or something along those lines.
Yeah.
Speaker 1 (09:25.538)
But if that didn't happen, then it's gonna be really hard for you to prove discrimination. And that's why failure of higher claims are oftentimes even harder to pursue. Unless you can get your hands on some statistics or something or something to show a trend.
I know that occasionally I've heard stories about studies that have been done where resumes are submitted and with all the details the same except for the name.
Yeah, and those studies generally do, the ones that I'm remembering, tend to show a significant difference in callbacks for interviews based on the name change. And the name change being designed to suggest either it's female versus male or that it's a particular race.
So this gets into one of the publications that I saw listed on your faculty page that I wanted to ask you about. You wrote an article on the topic of algorithmic affirmative action. before we even get into that and the substance of your argument about that, I'd just like to take a couple steps back to ask a little bit about the landscape here. So I know.
Personally, I've heard anecdotes for years about how people need to be more adroit when they're putting together their resumes nowadays, specifically because of the fact that they are probably going to be reviewed by a machine before they get reviewed by a person, right? And therefore you almost have to do almost something kind of like keyword optimization in order to make sure that a resume actually gets viewed by a person.
Speaker 2 (11:17.006)
So I'm interested
keyword optimization where you make sure it doesn't include certain words.
really? That's equally fascinating. I mean, number one, I'm just curious to hear what you have to say about that. But then also, I'd like to ask you just a little bit. Yeah, if you could just talk a little bit about what role sort of algorithms are starting to play in the hiring process.
Yeah, so I think the most obvious potential use and I think where they are practically most effective for employers are weeding through relatively quickly, huge volumes of applicants. So usually you're talking about applicants who are not current employees. You're not talking about applications for a promotion because that's a smaller body, but you're talking about identifying and recruiting candidates.
And sometimes it's the employer looking for the candidates based on things that are posted on the internet. But a lot of times it's weeding through the applications that were actually submitted or the resumes that were actually submitted. And I think that's where employers get the most benefit from the machine learning process, the most efficiency gains. But...
Speaker 1 (12:33.782)
The, what the, what the machine learning algorithms are attempting to do is predict who's going to be a good employee. And they have to do that based on examples that they have of people rated as good employees. So basically the machine learning algorithm is fed a whole bunch of data about actual employees. And then including data about how well they performed. And then it tries to draw correlations that, that humans can't really.
necessarily wouldn't necessarily be able to identify. And the main benefit, the main claimed benefit of machine learning is, yeah, it's really good at identifying correlations, even correlations that you wouldn't see if you were a person just trying to analyze or a human trying to analyze it. yeah, maybe it can uncover some really important factors in being a good employee that you might not think are factors in being a good employee.
But of course, the problem is when you're using examples from the real world to train the machine learning algorithm, that means you're feeding it examples of data that are painted by, potentially, by discrimination that happened in the real world already. So for example, if white males generally, because of the bias of the supervisor rating their performance,
white males tended to get higher scores than say females or minorities, then that data gets fed into the machine. And now it's predicting based on whether it can figure out if you're a woman, if it can figure out that you're a woman, it predicts you to get a low score. If it can figure out that you're a man, then it predicts you to get a high score and be a good employee. So bias comes out on the other side. Now the obvious sort of naive
thinking on a potential solution is, well, just don't tell it what the sex is, right? Just don't tell it what the race is. Don't tell the machine what the race is or what the sex is. The problem with that is it doesn't work because the machine learning algorithms are very good at identifying correlation. So they can basically find proxies for race and proxies for sex. So if you mention a particular kind of activity, gymnastics, for example, as a thing that you like to do.
Speaker 1 (14:55.864)
the machine will identify a correlation, say, that's probably a female and therefore they're probably not as good at performing. it will, without knowing the sex, it can still operate in a biased way.
So in that case, algorithmic affirmative action would be the attempt to sort of deliberately counteract the consequence of some of these machine learning algorithms that have this bias encoded within them. So how have people attempted to do this?
it.
Speaker 1 (15:31.522)
Well, so there's a lot of different ways to attempt to do it. Some sort of just try to do it by imposing kind of a mathematical constraint, like the output needs to have a selection rate of men that's within a certain range of the selection rate of women. So like a mathematical constraint on it. And that's actually a common way to try to do it. So.
the simplest example and it follows an EEOC guideline on this. The selection rate for females can't be any less than 80 percent of the selection rate for men. There's an EEOC rule on this called the four-fifths rule, the 80 percent rule.
It's an imperfect rule. Everyone recognizes it as an imperfect rule, but it's a rule of thumb. Well, you could build that kind of mathematical constraint into your machine learning algorithm and say, don't give me any output that fails that test. then you're fiddling with the output you would otherwise get from a pure machine learning algorithm that was unconstrained, which is why
arguably you're taking a sex or race conscious factor and injecting it into the algorithm. There's some much more complex, complicated ways you can try to build fair, what the machine learning scientists will call it fairness. There's complicated and complex ways you could try to build in fairness constraints into your machine learning algorithm. But in order to do it, you usually have to have the machine.
be aware of the sex or race of each of the applicants. It has to know who they are in order to employ the constraint. Arguably, the article talks about how that may be in tension with some US Supreme Court precedent from the non-machine learning context. The Supreme Court hasn't dealt with this at all. But it has dealt with situations where, for example, the leading case on this is a case involving
Speaker 2 (17:46.541)
Yeah.
Speaker 1 (17:55.084)
the city of New Haven and the hiring of firefighters there. They gave a written test to firefighters for promotion into higher level officer positions. They went to pretty good lengths to try to make sure that the result wouldn't be racially tainted.
an oral component of the exams as well. They went pretty far to try to make sure there was no racial tainting of the disparity in the output. But after they gave the test, came back to promotions were still mostly going to white firefighters and not to minority firefighters. The city wanted to throw out the test and not use it to make promotions. That was after a long debate, but the Supreme Court ultimately said,
it's disparate treatment. It's a violation of the statute to throw out the, and not use the results of the test for promotions just because of the racial disparity. You're actually throwing them out just because you didn't like that the results were skewed in favor of white firefighters. And so you at least need to think about whether or not you need to analyze whether the...
the procedure that you use or the test that you use, the selection device that you use, you need to consider your defenses to a claim that it caused a disparate impact. And so if that's the rule in a testing situation, the question is, how does that apply to a machine learning fairness device that's designed to produce output that is not biased?
intentionally injecting race or sex into the decision-making process in a way that arguably happened in that New Haven case. That's funny, when talk to machine learning scientists about that case, the Ritchie case, the firefighters case, they're surprised that, it's a violation of the law to
Speaker 1 (20:18.194)
throw out the results or attempt to adjust the results? Yeah, that's based on race or based on the racial output. And the answer is, yeah, it is. So then their question is, well, how do you do fairness in the machine learning context? And I was like, exactly. That's the point of my article. Can you do it? Can you do it?
That's the issue at hand. Yeah. So I realize this is a different context. This is talking about college admissions as opposed to employment. But I'm curious to know whether or not you think the decision that the Supreme Court made in the Harvard case last year about affirmative action in college admissions, whether or not that suggests to you at least like, whether that suggests something to you about the imminent direction of the court when it comes to more cases along these lines.
Absolutely. Great question. So my take on that is it didn't say anything about employment. So it doesn't affect Title VII, although one of the concurrences suggests strongly that they think it should. It was of course, it's just concurrence. He directly draws comparisons to the Title VII context and an opinion that he wrote in
which can pass.
Speaker 1 (21:37.462)
in the the Bostec opinion, the LGBT opinion that he wrote in the Title VII context. So he clearly sees the connection. He actually thinks, I think, that the same result should happen in Title VII. But the majority opinion doesn't say that. And so I think on its face, that recent higher education admissions opinion doesn't change employment law.
or employment discrimination law, but it certainly portends the possibility that the Supreme Court may do something very similar in the employment context in Title VII.
Another article that you wrote that I'd like to ask you a little bit about was you've written a little bit about OSHA. in particular, talked about actually before we get into the nuts and bolts of what you wrote about it. First, can we just talk a little bit about OSHA? Like, how long has OSHA been around? You know, what does it do? Like, let's let's talk a little bit about that.
So OSHA was, let me make sure I get the year right. OSHA was enacted in 19, I'm gonna say 1970, but I wanna see if I got it right. Yes, OSHA, the statute was signed 1970 and the,
It like these things always date back to the Nixon administration.
Speaker 1 (23:01.802)
Yeah, and it was bipartisan. OSHA basically is designed to provide federal floors for safety and health standards in the workplace.
Yeah, yeah.
Speaker 1 (23:28.138)
At its inception, idea was, we will basically imported a bunch of existing industrial standards at the time and made them have the force of law by becoming OSHA standards. The idea was, and will as science advances, in particular with our knowledge of exposures to substances, as science advances, we will enact new OSHA standards.
that will then become effective and we will continually update the standards in a way that keeps workplaces safe to the state of the art, The state of the scientific art, I should say. But the problem has been that there were some Supreme Court opinions that really made it difficult, it made the burden very high on OSHA to adopt new
OSHA standards, especially for exposures to substances, chemicals and substances used in the workplace. What's happened is basically we get almost no new standards ever on substances and chemicals in the workplace from those that were adopted right after OSHA's enactment in 1970. Our standards basically, there are a few exceptions.
There are a few substances for which we've updated the standards or tightened the standards and made the volume of the substance in the air or the permissible exposure limit tighter. hexavalent chromium is one of those that we... After years, it's usually like years or potentially decades of fighting, maybe we get one standard updated.
So what's happened is the vast, vast majority of all the standards that are enforceable, particularly on substances and chemicals in the workplace, they all date back to like the 70s. Standards that were enforced in the 70s, which even kind of date the science dated back before that. it's kind of stagnated, particularly the development of new standards has really stagnated.
Speaker 1 (25:53.354)
It's so much so that the former director of OSHA basically issued a public call requesting advice and saying, are there any ideas for how we can do this in a way that's consistent with the Supreme Court has said? Can we update standards in a more efficient, faster way? We haven't seen that. Now, OSHA doesn't just regulate
hazardous substances and chemicals, that's what I focused on a little bit more. But there are also workplace construction industry standards, things like that, like scaffolding requirements and those things. There's also a general duty clause in OSHA, which basically provides that employers have a general duty to provide a workplace free from known risks.
known risks that could cause serious injury or death.
So could you tell me a little bit about the particular article that you wrote about this? And so I was looking at the abstract for the article and let me give you my take on what it sounded to me like you might've been arguing, but I'd be curious to know whether or not. So it sounded like you were saying at first that people have occasionally made the argument against OSHA requirements in the first place to say that.
they're largely unnecessary because the market can effectively address any problems that would arise from workplace safety, right? That eventually the market will respond to it in some way.
Speaker 1 (27:34.84)
An employer who doesn't provide a safe workplace, in theory, would get punished by making it harder to hire employees because they'd have a reputation for putting their employers in danger.
Yeah. so it's not to me like you were saying that countering that you were saying that OSHA is still necessary to help ensure some basic level of trust and an assumption of good faith between employees and employers, because the risk is too strong that employers would otherwise just take advantage of the total lack of regulation to opportunistically exploit employees for brief periods of time before word got out about what they were doing.
Is that? Yeah.
Yeah, that's I mean, that's basically right. The other concern is that. If you think about how the order of performance of the employment trade off, so an employer hires you, you show up for work and you do a good job and you keep doing a good job and you get paid. That's that's the basic agreement, right? The the the problem is that if you get hurt such that you.
can't be an employee maybe. Maybe you're permanently disabled or permanently totally disabled. You can't be an employee. Now the employer has this incentive to reduce whatever you recover. they have a chance, if totally unregulated, they have a chance to take advantage of the order of performance because they got all the great work from you while you were good employee. And now that you're hurt, well, we can...
Speaker 1 (29:15.714)
We want to try to deny you workers' coverage on the backside of that, right? Or we will claim that it was actually your fault or that you were disobeying the policy. We instructed you not to do it that way now because they have every incentive to do that at this point to claim it's not a compensable injury so that their insurance premiums go up less and so on. So the structure of the workplace dynamic makes it so that there's
this sort of opportunity for employers to take advantage of the timing of performance. And one of the things that OSHA can do, can be effective in, is forcing the employer on the front end to actually be safe, right? To do the things necessary to protect the safety and health of their employees.
So to give you an example that's in the article from a really old case, there was a workplace rule that employees weren't supposed to ride on a particular vehicle when they were at work. But you'd get to the job site quicker and faster if you did. So the policy was you can't ride on that particular car.
employees did it all the time so that they could get to the workplace and the employer didn't punish them for violating that policy, right? Because it actually was better for them. The employee got to the workplace faster, it was more efficient, right? Until of course, someone rides on the cart and gets hurt. And then the employer says, yeah, but you shouldn't have been riding on that. That was a violation of our policy. So that's an example of how an employer can take advantage of the timing of performance, right?
Yeah, you were violating the workplace safety policy, but we didn't care because it was more efficient. It was faster. You got the work done. But then when you got hurt, the answer is, yeah, well, you shouldn't have been doing that. That was your fault. And so OSHA can, the front end regulation that OSHA does could, if it's effective, reduce that temptation to act opportunistically by the employer.
Speaker 1 (31:39.778)
They could say, we can't let you ride in that particular car. It's our policy because OSHA says we can't do it. It's unsafe. so that's one role that OSHA can play. It can prevent employers from engaging in that kind of opportunistic taking advantage of the timing of performance.
Okay, interesting. So I'd to ask you a little bit about, obviously there's been a lot of discussion about AI recently and about sort of what kind of big changes in the employment landscape are on the horizon as a consequence of AI. And we were just talking a little bit about algorithms and machine learning as they apply to
the job application process. But I'm curious to ask you just a little bit about what you think about sort of the future of white collar work on the basis of the rapid development of AI. Not to ask you about something way too broad, but.
As I'm looking at the zoom screen that has the AI companion button at the bottom.
Yes, I know they certainly try to get you to adapt at every possible opportunity.
Speaker 1 (32:57.846)
I think it, mean, I think you're probably talking about for the most part, things like generative, text generative AI, right? So, so chat GPT and those sorts of things, similar types of tools. I think it's, you know, as I think about it's, it's natural for me as a lawyer to think about the white collar work that my graduates will go do or that I used to do.
and think about how generative AI would potentially change that landscape. Because generative AI like that can be a really valuable tool in some ways, and can really provide some efficiencies that are useful, but only if they're in the hands of someone who knows and is careful about the limits of AI and recognizes, OK, well, what is this actually doing? It's just
finding the average most likely next few words to be used, right? So you're not actually talking to a computer, you're not talking to something sentient that's thinking about and then responding carefully to your answer. It's just using correlations to come up with what it thinks is probably the answer. Or I shouldn't even say what it thinks is probably the answer. It's using a correlation to spit out the most likely next words in response to your query.
So.
You know, when you think about what lawyers could do with that, well, you could quickly write briefs that way, right? even more, the problem of course is if it's hallucinating case names that aren't real cases, or if it's wrong, right? It's citing a case that's wrong or it's been overturned or something. Like if you're not careful with that, you're on the cusp of committing malpractice by using it.
Speaker 1 (34:59.114)
On the other hand, you can imagine that it would do a good job of giving you something to start with. It will give you the bare bones outline of something. And you could carefully go through and change and correct things. in that way, a well-informed and careful user of it could make it into a pretty efficient tool. I was thinking about the
The limits of this though, at some point, it seems like the legal profession is going to have to think about where to draw a line. Imagine that an employer and that a lawyer walked into court for an oral argument and they just had chat GPT going and then they could respond, they could just input the judge's question via chat GPT on their phone.
have it listen to the judge, and then it would spit out the correlated answer, right? And then the lawyer could just, you could actually just use chat GPT as opposed to engage in your oral argument. Now that seems pretty far-fetched, but you know, at some point the legal profession's gonna have to say, this use of it's allowed, this use of it's not allowed. I don't know exactly where that line is, but if I was a student today,
I'd want to know everything I can about chat GPT and its limits so that can know how I can use it and how I should and what I need to be aware of the pitfalls of using it, potential dangers of using it. I think it's something that everybody in a profession like the law should be thinking about and nobody knows exactly how that's gonna play out.
Yeah, I mean, I know I'm asking you to look into the future here, but like, do you share some of the concerns that have been expressed about like the possibility that this is going to result in massive job losses, not necessarily just in law, but just like throughout the, you know, throughout various white collar fields, or I know people more optimistically say that they think that, well, this is just the latest technological development and people will, it'll change the way.
Speaker 2 (37:20.492)
work looks, but it's not going to supplant it.
I'm probably in the latter camp on that. I think it's probably going to change the way people do certain kinds of work, but there's still going to be great value in the human resource of the human judgment that's brought to using that technological tool. Computers didn't put us all out of work. They very much changed how we work.
Email didn't put us all out of work. It changed what we do every day. It changed how we work. But I think it's a very powerful tool that people are going to have to figure out what it means for their particular industry. And it could cause some uncomfortable shifts in...
available positions, right? I mean, there aren't, you know, you can think back before computers to jobs that existed like in the legal field, you would have a legal assistant taking your dictation notes, right? The partner would sit and dictate whatever they were going to write, a brief or whatever. They would dictate it.
and there's somebody whose main job was just translating dictation and then going back and forth with edits. They don't do that so much now, but they do still manage technology, just different technology, and they help the lawyers. So I think it's sort of like that. It's gonna be a new tool. It's gonna change the way some people do their jobs. It may change the availability of certain jobs, but I don't think it's gonna put all lawyers out of work.
Speaker 1 (39:19.35)
I put all...
what paralegals out of work, whatever, or all accountants out of work, whatever. I don't think it's gonna have that kind of effect that leaves whole swaths of people without the ability to support themselves and without work.
So I'd like to ask you a little bit about what do you think are some of the major issues in employment and labor law that are on the horizon and in particular, like working their way up to the Supreme Court?
So this one's less working its way up to the Supreme Court, but I think it's an area that's going to develop probably fairly rapidly over the next decade or so. And that is data privacy and how data privacy regulation applies or doesn't apply in the employment context. In Europe, the GDPR
is applicable to employees and employers and what employers can do with and about the data of employees. In the United States, we don't have that yet, but California has a data privacy law that only recently became now, it was originally only applicable to consumers, so consumer data, right? So you agree to,
Speaker 1 (40:53.386)
sign up for an email list or something, and then the company gets data about you, or you go shopping at a grocery store and you give them some data. There's some limits on what, many states have limits now on what entities can do with consumer data. But employee data is different, and most of those laws don't touch on employee data. But California's law,
has recently been made applicable to employees. So more like the GDPR where there is data privacy restrictions on the use of employee data. I think that that area is ripe for potential change in the United States. I think as people become a little bit more aware or appreciate a little bit more the amount and types of data that's collected.
about them or on them or their data in the United States. As employees, think there may be some bipartisan pushes to include restrictions that look more like the GDPR. That could potentially have very significant ramifications for employers in the United States. They'll have to change the way they do some things.
One example of that is Illinois has a biometric data privacy act. And a couple of players have been hit with some really massive judgments just because they were fingerprinting employees as like time check-in, right? So like they fingerprint to be, they have a fingerprint in order to enter the workplace, right? Or fingerprint time punch cards.
they didn't get consent for that. And so they were in violation of the Illinois biometric statute and they got hit with a huge judgment. And that's the sort of thing like employers will have to learn all those laws in all those different States and make sure they're in compliance. And so I feel like there's a potential for development by state laws or in theory also possibly federal law in the United States, but there's potential for development in state laws there. And if we get,
Speaker 1 (43:16.8)
a really varying patchwork across the states of data privacy restrictions on what employers can do with employee data, then employers are going to have to spend a decent amount of money on attorneys to try to figure out what they can and can't do with their data collection.
Do you think that California is going to kind of lead the way when it comes to some of this stuff that like, it's going to be the sort of situation where whatever California adopts ends up becoming the new national standard or?
I don't necessarily think that because California is actually kind of following the European standard. So the question is really, are more states going to weigh the costs and benefits of this and go, yeah, we actually want to be a little more protective like Europe, or are they going to take the approach that the more hands off approach to that. And it's interesting that you kind of actually do get by part. This is not an issue in employment law where
Democrats are always on one side and Republicans are always on the other side. This is an issue where people are like, across party lines, people are like, yeah, well, maybe we need to be a little more protective of individuals' or a little more, need to have a little more disclosure about what data is taken and how it's used and what you can do to correct it it's wrong and that sort of things. Using.
Having to use your fingerprint in order to clock in to your job is, that's getting into like quasi-Gatika kind of territory.
Speaker 1 (44:44.544)
Yeah, or any other biometric data that's like imprinted on your ID card. Sometimes security systems will be such that they use some kind of biometric data marker about you that's matched on the card to whatever in the system. And so when you swipe your card, it's verifying the two things. And it uses biometric data to do that.
And employers might think, this is cool technology. We should use this. But they need to be thinking about the potential that it's being regular.
Yeah, yeah. Other issues that you see kind of like on the horizon? Big issues?
Well, currently in front of the Supreme Court, there's a question of whether a lateral, well, whether a transfer from one department to another department that doesn't have any significant disadvantages, right? There's not, you don't take a pay cut. You don't have fewer opportunities for overtime. There's no finding that it's disadvantageous transfer. It's just a transfer from one department to another.
whether that is enough of an action to support a Title VII claim, like whether or not you can bring a discrimination claim based on a lateral transfer.
Speaker 2 (46:00.276)
interesting. I never thought of that.
And because a lot of the courts had developed some kinds of limiting tests on what amounts to a materially adverse action or an action that is bad enough that you can sue, right? And the courts, some of the thinking there as well, you can't sue for anything bad that happens to your work. It's gotta be something significant. Like you get moved from one office over to the other office across the hall and it's a better or worse view. Can you bring a Title VII claim based on that?
And so, well, that's kind of the question that's, the court has kind of narrowed the question presented, but that's the question that the court's gonna have to consider here because the statute doesn't contain any sort of limiting language like that. It just says discrimination in the terms, conditions, or privileges of employment. And of course, like which office you're in, I suppose, is a term or condition of your employment. And if they move you from.
an office with a better view to an office with a worse view and you say it's because of sex or because of race, is that a violation of Title VII?
Huh, that's fascinating.
Speaker 1 (47:10.254)
So that one's in front of the court right now. Other issues that I think may make their way up to the court, there's long been an issue percolating about the standard for proving discrimination in the ADA disability context. So the court has given us some, well, let me back up a little bit.
Title VII uses the key words there because of, so you can't discriminate in the terms, conditions, or privileges of employment because of somebody's in Title VII race, color, sex, national origin, or religion. And similar language is used in the age discrimination context in the ADEA, the age discrimination statute.
It also says because of, but those have been interpreted differently. So under Title VII, you could have a mixed motive case and you could still have a violation of Title VII. So if your argument was, was, maybe the employer didn't give me the promotion because of my performance scores, but also part of the reason, one of the motivating factors was sex. Well, that would be a violation of Title VII. But in the age context,
Age has to be the but for causation. In other words, you can't use a mixed motive theory. You can't say it was age plus my performance. That wouldn't be a violation of the statute. You have to prove as plaintiff that if you were younger, you would have gotten the promotion. In other words, age was not the but for cause. The court hasn't answered the question of what's the standard for disability discrimination. That's a separate statute, the ADA.
I anticipate that at some point somewhere along the way, the court's going to answer that question because they've done it. They've answered this question for the age discrimination statute. They've answered it for section 1981, which is a reconstruction era race discrimination statute. They've answered it in the retaliation context. And they've said there in retaliation context, it needs to be but for, even though it's also Title VII.
Speaker 1 (49:28.43)
So they've had several cases answering this question about what is the causation standard under each statute, but they haven't answered it yet for disability. So I anticipate that that one will eventually make its way up to the court.
could you, could you talk a little bit about what you mentioned? Bostock earlier. Could you talk a little bit about Bostock versus Clayton County and, how that I, my understanding is that it, it, to some extent, it expanded the entitle of, the interpretation of title seven. Would you say that's the case or.
Yeah, I think that's fair because if you look back, you don't have to look back that far. If you look back about a decade, there were no courts that were saying that because of sex in Title VII included discrimination because of sexual orientation or gender identity. But relatively recently, a few circuit courts had decided even on bank circuit courts, so the full court, the Second Circuit and Seventh Circuit,
had decided that actually, logically, you can't discriminate against someone because of their sexual orientation without considering sex as one of the factors. In other words, if you fire a male employee because that person's significant other is also male, that means you did it because of sex, because if that person were a female instead and had the same partner, male,
then they would have been heterosexual and you wouldn't have fired them. So sex is a factor in those cases. And so a couple of circuit courts went that way, as I mentioned, the second circuit and the seventh circuit went that way. The 11th held to the contrary. so it made its way to the Supreme Court and the Supreme Court answered the question. And I think some...
Speaker 1 (51:27.16)
people might've been surprised by that result. The result being that sex, because of sex in Title VII does include sexual orientation or gender identity. So it is now a violation of Title VII, the Supreme Court has held a violation of Title VII to take an adverse employment action against someone because of their sexual orientation. Now.
If you'd asked people in 1964 when Title VII was passed, if that were true, one of the arguments in this case was nobody or no legislature, no legislators would have thought that when they said because of sex, they were also including because of sexual orientation, because people tend to think of that as a separate category. So yeah, in that sense, it expanded.
the reach of Title VII because courts weren't applying it that way until very recently. And the Supreme Court now says that it is a violation of Title VII. The most interesting thing to me about that opinion was that there were three different opinions and all three of them claimed to be faithfully applying the ordinary meaning of the words because of sex in Title VII, which can't possibly be true, right?
Someone needs to be right about this, but they all were taking the textualist approach. They all said the text compels this result. Justice Gorsuch wrote the majority opinion. says it necessarily, includes logically, it includes sexual orientation. Even if people didn't think that it included sexual orientation, the words they chose actually did include sexual orientation. Justice Kavanaugh writes a dissent.
And he says he's taking the textualist approach as well. But he says in the ordinary parlance, people thought of sex as something different than sexual orientation. And so they included sex as a thing that you can't discriminate on, but they didn't include sexual orientation. So they made a choice between those two, essentially. So everyone claims to be a textualist, but they reach totally different results.
Speaker 2 (53:44.908)
What was the third interpretation?
Well, the third interpretation, the third opinion agreed that sexual orientation was not compelled by the text. That was Justice Thomas and Alito and that opinion. So they agreed with Kavanaugh, but they wrote separately.
Okay, okay. So I wanted to ask you about Proposition 22 in California back from 2020, which was about how tech companies classify workers, whether or not they count as independent contractors or as full-time employees. And given what an explosion there's been in the gig economy and people who work for Uber, Lyft, DoorDash, Instacart, all those sorts of...
companies. And I'm curious to ask a little bit about your your opinion about that and how that's been playing out.
Well, so this is classic problem in employment law. The app-based gigs are a new version of it, but there's a classic problem in employment law. Who gets protected by the statutes and who doesn't? Who's an employee and who's not? Now, usually it's to your benefit to be an employee because you get the statutory protection, the statutory coverage.
Speaker 1 (55:09.442)
But there's some downsides to it. One downside is workers' comp. If you get hurt and you have a tort claim, if you're an, ordinarily you'd have a tort claim, but if you're an employee, as opposed to an independent contractor, then you're probably limited to workers' comp recovery and you don't get to bring a tort claim. So sometimes you'll see on the flip side, you'll see,
you'll see workers arguing that they were actually independent contractors, not employees. So you will see that sometimes, but usually there's protections provided by various state and local statutes and federal, and it matters whether you're an employee, if you're an employee or protected, you're an independent contractor, you're not. I think what happened in California appears to be a...
it's posed as a compromise. So the idea is, hey, we're gonna let you, you're gonna be classified as independent contractors. So you're not gonna get statutory protection that are the other employment statutes, but we're gonna build in some minimum floor things for these gig workers. And those minimum floors will protect you instead of the employment law. And that is to your benefit employee or worker, I should say. That is to your benefit worker because.
benefit worker Uber driver, because you then have total control. Like you're an independent contractor, you can do what you want. You can set your own schedule and so on. I think it's an interesting compromise, certainly pushed hard by Uber and Lyft and the app-based services, the gig-based services, delivery services, pushed hard because...
To them, it's a huge chunk of their business model that these people are independent contractors and not employees. We don't have to meet all of the requirements of various state laws protecting employees. And this was a voter pass initiative, but heavily pushed by the tech apps. And I think it's in front of the California Supreme Court now under the California Constitu-
Speaker 2 (57:23.711)
really?
the California constitution, question whether it violates the California state constitution. It could potentially serve as a model for other states to think about. It's essentially like creating a third category. You're not just an independent contractor, you're not an employee, you're this third thing covered by this separate law where we build in a nice floor for you, but you don't get all the regular.
benefits of being an employee covered by all the other employment statutes.
What was that floor that was built in in that kind of compromised position that was represented by the proposition? Because I don't think I knew about that.
There's like a wages floor and a health benefits floor. So there's some obligations on Uber and Lyft and the app providers that you wouldn't have with any other independent contractor.
Speaker 2 (58:12.416)
Okay.
Speaker 2 (58:22.089)
I see. Okay, interesting.
So it's kind of like creates a third category somewhere in between with these different protections. But one of the challenges is workers comp because one of the challenges in that case that's in front of the California Supreme Court is workers comp because once you classify them as independent contractors, you pull them out of the workers comp statute, then arguably, right, you're running a foul of state law or
Okay.
Speaker 1 (58:53.742)
the state constitution, arguably, on the trade-off between tort and workers' comp. So maybe if a state's thinking about creating this third category, they also got to think about what happens if the employee gets injured. Do they help tort claim? Is it the tort claim barred? Do they get workers' comp, access to workers' comp? What's the answer going to be on that?
Could you just talk a little bit about the employment and labor law externships at Stetson?
Yes, happy to do that. So, and I'm glad you brought that up. So we have a labor and employment law externship that we've been doing for, guess, we've had an EOC placement for a long, time, but we've also got a couple of new placements. So we now place externs at the Equal Employment Opportunity Commission, which is employment discrimination. They administer the Title VII,
the ADEA, Age Discrimination Employment Act, and the ADA, the Americans with Disabilities Act, and some other statutes like Genetic Information Non-Disclosure Act. We place externs at OSHA, and the ones at OSHA, they're actually not doing primarily safety and health work. They're actually doing whistleblowing, investigating whistleblower complaints, because OSHA has given
sort of a weird structure, but OSHA has been given the authority to administer whistleblowing provisions of lots of different federal laws. And so mostly what our externs are doing at the OSHA placement is working on and investigating these whistleblower complaints. So it's also kind of like a discrimination claim, right? Because you've been fired because you blew the whistle on an agency for whatever reason.
Speaker 1 (01:00:50.606)
or an employee of an agency, something like that. So it does kind of look a little bit like the, like discrimination work, even though it's at OSHA. And then the NLRB is the third place where we place externs. And the NLRB is the independent agency that administers the National Labor Relations Act. So that really deals with organized labor, collective bargaining and what
is an unfair labor practice or what is or is not an unfair labor practice by a union or by an employer in response to potential unionization activity. the, the experts that we placed there are generally looking into charges of unfair labor practices. They also, if there's an election going on, they might get to observe the election, but that, you know, that's kind of hit or miss depending on the timing. So we have experts at all the three of those placements.
And you can do those externships either during a fall or spring semester, or you can do it during the summer session. So we usually can place one or two students at each of those placements. We currently have two at EEOC and two at OSHA. And we don't have anyone at NLRB this semester, but often we do.
I'm out.
Speaker 2 (01:02:12.056)
That sounds like a great opportunity.
It's a really good chance for students to get to see sort of behind the scenes at the agency, see the work that goes on, also get familiar with the substantive law and see how it works in practice. And so it's really a cool opportunity. we've we've had externs go on to, you know, successful careers in labor and employment field. And actually one of our Stetson grads recently started as an attorney at the EOC. So it's great to see that when that happens.
Yeah, yeah. All right, well great. Well, thank you so much. This has been Real Cases. Thank you for listening. Check back for more episodes about an array of legal topics presented by the Stetson University College of Law. Learn more at stetson.edu.
Happy talk.
Topics: Real Cases Podcast