Season 3, Episode 5: AI in the Workplace: Risks, Regulations, and What Employers Need to Know

Daniel Schwartz: So, I'm going to assume at this point that most podcast listeners have at least heard of sites like ChatGPT and other generative AI websites. And I know we have a discussion just about that, but I heard this thing recently that described it in this way - and I thought we at least start the discussion with it - which is that what gen-AI does is to figure out what the best next word is, but to do so in things like sentences and paragraphs. So, if I said Hartford is the capital of the state of and I left it blank and asked you to figure out what the next word was going to be in that sentence, I think most of the listeners would likely guess it would be Connecticut. For those who thought it was the capital of Massachusetts, I'm sorry, I can't help you there. But that's in essence what the program is designed to do. And it doesn't think on its own, at least not yet. But there are other AI tools out there as well. So, Emily, maybe let's start from that perspective. How are employers really using these AI tools in the workplace?

Emily McDonough Souza: Yeah, they are using them in a ton of different ways. But just to name a few, we're seeing these tools be used for electronic monitoring and productivity monitoring. And what that means is that employers, for example, are placing software oftentimes called bossware on an employee's device to, for example, track their mouse activity, track their keystrokes. This type of software can even track how long an employee is looking at their computer screen versus looking elsewhere in the room. And we're also seeing other types of monitoring tools using AI, such as cameras, for example, in delivery trucks. So, in the case of Amazon we're seeing that Amazon is using a four part camera system with biometric feedback indicators. And what that means is that it's monitoring if the driver is looking away from the road, if the driver is speeding, if he's yawning, he or she is yawning, stops at a stop sign, puts on his or her seatbelt, and then the camera can also send live feeds to managers to enable them to watch the drivers as well. So, these are just some examples of electronic monitoring and productivity monitoring that's being used in the workplace.

Daniel Schwartz: Wow. And so, I would suppose one of the other areas that employers are using AI is going to be in the hiring process, maybe to analyze resumes or things like that.

Emily McDonough Souza: Yeah, absolutely. And I think it's important to know that there are a lot of pros and cons to using the tool in this type of context. But essentially what happens is that AI can be used in recruiting and hiring new employees by preparing job descriptions and in screening resumes to identify top applicants for certain positions instead of having, a human do it.

Daniel Schwartz: Yeah, I suppose in any area that has perhaps a lot of data that needs to be analyzed quickly, that's where AI is going to be useful. And I think that we've all gotten now used to some AI in our lives. If we've ever watched Netflix, Netflix has some AI or some rules based program that's going to recommend something else that I might want to watch or Amazon might think that, hey, if you've bought X, you might also buy Y. Yes, it's not that surprising to see it in the workplace, but I think the speed in which it's doing so is really taking some by surprise. We thought maybe we'd have a little bit more time, right?

Emily McDonough Souza: Right, yeah.

Daniel Schwartz: Yeah. So, all right. So, we've talked about some of the areas where employers are using these tools. What are some of the risks, though, that employers need to be aware of? I suppose some of the existing laws, do we need to still worry about that?

Emily McDonough Souza: Yeah, absolutely. And, as you said, this is such a quickly evolving, new thing in the workplace that it's important to remember that employers need to comply with these existing laws. So, in terms of what laws we're talking about, we're talking about Title VII, the federal law that prohibits employment discrimination based on race, color, religion, sex, national origin. We're also talking about the ADA, which of course prohibits discrimination based on disability. An example of this would be in the AI context where we're seeing if a company uses video interviewing software that evaluates candidates based on their facial expressions, based on their speech patterns, tone of voice, to determine if they should receive a follow up interview. The program might you know, unintentionally screen out disabled applicants who have speech impediments, who are deaf or blind, have had a stroke, especially since most of those platforms are designed using a sample of able-bodied testers. Another law we have to be aware of course, is the ADEA, which prohibits discrimination based on age. And then wage and hour laws. For example, we have had reports of some employers using automated activity to track and then deduct money from remote employees paychecks because they are detected as not looking at the screen for a certain amount of time, a chunk of time, then the employer might claim that they were not working during the time detected by the AI. And of course, as you can imagine, the DOL has stressed that relying solely on this automated timekeeping and monitoring systems without human oversight can run afoul of wage and hour laws. So, something to definitely be aware of.

Daniel Schwartz: Yeah, and I think that brings up a great point. If an employer is doing that, maybe doing it improperly, having it being done on a company wide basis could really subject the employers to a class wide claim. I would think that employers really need to be mindful about implementing these tools, because if you're doing it on a class wide basis and you may be violating the law, that can create big exposure to clients.

Emily McDonough Souza: Absolutely.

Daniel Schwartz: So, we have seen, I think, bubbling up, some other rules, but there haven't been a lot of nationwide rules, certainly. And there are only, I think, a handful of either states or local jurisdictions that talk about it. One of them is the New York City rule on automated employment decisions, right?

Emily McDonough Souza: That's right. Yeah. New York passed a rule about, like you said, the use of automated employment decisions a couple years ago it went into effect last year, and essentially it prohibits employers and employment agencies from using these AI tools unless they conduct a bias audit on the tool within the previous year. They have to make the results of the bias audit publicly available on their website. They need to also notify job candidates and employees that this automated tool will be used. And a bunch of other parameters as well - provide instructions for requesting reasonable accommodation, for example. Use an independent auditor who is capable of exercising objective and impartial judgment. And, they are levying, you could have fines levied against you, if you are not in compliance with this New York City law.

Daniel Schwartz: I think that's probably a good point. We're not going to get into all the laws on this podcast, but Illinois had an artificial intelligence video interview act and I know in Colorado. Back last May, they had some tort liability for AI Algorithmic discrimination and employment, and some of those laws may go into effect in early next year as well. So, I think In light of some of the disruption we've had in D. C., it's probably unlikely we're going to see federal legislation, so at least in the near future, probably makes more sense to look at state laws here.

Alright, so we've talked about how employers are using these tools and what risks employers should be aware of. What are some steps employers can be thinking about, though? Because they may want to use artificial intelligence in their workplace. They can still do so, right?

Emily McDonough Souza: Absolutely, Dan. And I think it's a commonsense perspective that an employer should be taking when they're evaluating whether and how they can use these AI tools. Instead of just, accepting a new product from a vendor that's on the market and putting it to use, they really need to review the products, take the time to really get behind the curtain and see what the product is, how it's used, evaluate any risks and make sure that, once they put it into use that they're not just washing their hands of it - that they are, doing checks and balances along the way, making sure that the tool is utilized in the way it's supposed to be. And they also should be, of course, utilizing AI policies with regard to employee handbooks and on their intranet. So, you know, this is the same thing that we dealt with when social media really first came into play, where policies, new policies, had to be designed to encompass these new risks and to provide guidelines for employees. So, I definitely think that is also needed here in the new world of AI that we're living in.

Daniel Schwartz: Yeah, that's a great point because I think employees are using this technology, whether employers want them to or not, and new policies or new guidance sort of instructing employees on the use of it probably an easy first step that employers can do to at least try to minimize some of the risk that employers might have because of the employees using it. We've heard stories of employees who are uploading confidential information. There have been stories of lawyers who have been using generative AI and getting hallucinations back on cases. So I, I think the primary focus for employers is, what the risk is for the employer using this AI, but I think you bring up a good point that employers also need to be thinking about how employees are using this and how can we regulate the usage in a way that minimizes the risk without stifling innovation, right?

Emily McDonough Souza: And then, of course, what we attorneys always say the other thing that employers should be doing is keeping up to date on the law, whether that be, getting counsel, certainly here at Shipman, you especially are always reporting on your blog about new developments in the law, but certainly keeping up to date on the law as this does have the potential to be expanding in different ways as AI becomes just a part of every day of everyday life.

Daniel Schwartz: Yeah, it's really amazing. When I went to this conference in November, the amount of discussion on artificial intelligence, you, you wouldn't have even seen this three or four years ago, much less ten years ago. And now I was struck by just the amount of programs just focused on this limited aspect of artificial intelligence.

Emily McDonough Souza Right, absolutely.

Daniel Schwartz: We were talking beforehand on some of the things that I gleaned from the conference, but I, I think I'd focus on three things that I thought may be worth mentioning. I hear one is a point you were bringing up to delve into deeper, which is there are starting to see some cases now that are coming out. We had this northern district of California case, this Workday, Inc. case that recently allowed a disparate impact discrimination claim to proceed against a third party HR software vendor who was using AI in job applicant screening tools, and it was based on the argument that the vendor was an agent of the employer to which the job applicant applied. So that was, I think, an important lesson that the existing laws still have teeth. That was something that was definitely brought up.

I think the second point that was brought up was the interplay between AI and then maybe employees right to either unionize or to discuss working conditions. So, to the extent that the employer was using AI to track where employees were going in the workday, maybe to see if they were congregating together, that it could have a an impact on the ability of employees to unionize or to discuss the terms and conditions of their workplace, which is managed by the National Labor Relations Act. So, that was an interesting discussion and one that I'm not sure gets thought about enough by employers -even those that don't have unions right now. And then I think the third takeaway that I had from it is, you can always learn more the technology. The AI that we have now is pretty much the worst AI is going to be in the future that the developments are happening at a rapid pace.

And it might be easy to discount the AI software right now, but maybe not a year from now, maybe two years, five years. We're going to really see some rapid changes and probably not in the way that we think about. So interesting points and all worthy of discussion. Emily, I'll give you the last word.

Anything else that we ought to be thinking about here?

Emily McDonough Souza: I think, like I said before, it's a common sense perspective, pros and cons to these AI tools that, that employers have at their disposal. But I think, just keeping ahead of any risks, seeking out counsel, keeping up to date on these laws, making sure that policies are in compliance, and they're writing new policies for employees, all of these things, keep in mind going forward, and yeah, it's a brave new world out there.

Daniel Schwartz: It is, I think anyone who hasn't explored some of these tools, it’s really pretty amazing. I was using one of the Facebook meta programs to create pictures, and it's really pretty remarkable in the last year, the developments of these AI tools to come up with pictures out of whole cloth. And I think we're going to see videos and other things that are either getting produced by employees or employers and really presents both an exciting time, but also one that might be hard for employers to manage.

So, I'm sure we will be back to this topic at another point in time just because the developments are happening at such a rapid pace. So, Emily, thanks for joining us.

Emily McDonough Souza: Thanks for having me, Dan.

Daniel Schwartz: And with that, we will close out another episode of From Lawyer to Employer. Again, we're going to come out with new episodes every few weeks, so look for them in your stream.

If you haven't already subscribed to the podcast, you can do so wherever you listen to your podcasts, and feel free to rate us or provide us a comment. If you'd like to share an idea for an upcoming program, feel free to email me at dschwartz@goodwin.com. Always love hearing from listeners of the podcast and readers of the blog as well. With that, we will close out another episode. Thanks again for joining us.

Host: Thank you for joining us on this episode of From Lawyer to Employer, a Shipman Podcast. This podcast is produced and copyrighted by Shipman Goodwin, LLP. All rights reserved. The contents of this communication are intended for informational purposes only and are not intended or should not be construed as legal advice.

This may be deemed advertising under certain state laws. Subscribe to our podcast on Spotify, Apple Podcasts, Google Podcasts, or wherever you listen. We hope you will join us again.