DE Talk

2026 Expert Compliance Insights: New Rules, Changed Government Enforcement & How Employers Can Respond

DirectEmployers Association Season 7 Episode 6

Federal contractor compliance underwent seismic changes in 2025: Executive Order 11246 was revoked, EEOC and the U.S. Department of Justice launched intensive investigations into DEI programs and EEO compliance, and the White House moved to centralize the regulation of artificial intelligence on the federal level–just to name a few.

As 2026 begins, employers need clarity on what's changing and what’s required. Employment law attorneys Mickey Silberman and Michelle Duncan from Silberman Law break down the essential compliance requirements federal contractors must understand and will provide actionable guidance on what you should do to ensure compliance and reduce risk.

DirectEmployers Association (00:00):

Get ready! The DE Talk podcast starts now. Insightful conversations and dialogue helping you put the human factor back in HR.

Jeff Gill (00:09):

Welcome to the DE Podcast. I'm Jeff Gill, direct employer's executive director, and I'm very excited to be stepping in today as today's host to discuss a topic that truly hits home with so many of our listeners. What to expect in 2026 with regard to compliance obligations, changing federal and state enforcement, and what may be next. We're very fortunate today to have two industry experts with us, Mickey Silberman and Michelle Duncan from Silberman Law to give us all the details. With so many changes in legislation in the marketplace directed at federal contractors, this enlightening conversation will help you prioritize what's important and offer practical guidance on what you should know and do to comply. Mickey and Michelle, thank you both for joining us today. This is a first for you joining.

Michelle Duncan (00:56):

Thank you for having us.

Jeff Gill (00:58):

To help our audience understand the depth of expertise that you both bring to this conversation, I'd like to spend a little minute with you sharing a little bit about your backgrounds.

Michelle Duncan (01:07):

Yeah, so maybe I'll take the first swing at that. So I was just looking at the calendar and this year will be 28 years that I've been working in the non-discrimination, EEO, OFCCP space. 28 years. I spent the first 14 of those working as a trial attorney for the US Department of Labor, where I represented OFCCP. So I was the attorney bringing cases against federal contractors. And then I'm almost to the halfway mark of my career, half of the time spent now representing employers. So I've been on both sides of the fence and that is in fact how Mickey and I met. When we met, I was a trial attorney representing OFCCP and Mickey was part of the defense bar and where he was representing federal contractors. So we got to go toe to toe on several different matters over the years, and that's how we became friends and how he wooed me away from my comfy federal government job to join the private sector.

(02:14):

So super excited, Jeff, to be here. And I work with a lot of employers who use DirectEmployers. I've hosted DirectEmployers folks at the Colorado ILG. So even though it's my first time doing a podcast with you all, I'm very well steeped in the good work that you do.

Jeff Gill (02:32):

Fantastic. Thanks for joining us today. Mickey, the team at Silberman Law is growing rapidly. Can you share a little bit about your firm and your area of expertise? Tell us all about Silberman Law.

Mickey Silberman (02:42):

Sure. So like Michelle, I've been doing this a long time. Shell, I'm not one upping you. 28 years might sound better than 31 years, but 31 years doing this stuff. By the way, I do have to give this real quick one, Jeff. As Michelle and I faced each other across the table on a number of matters over the years, I would go to her at the end of each one of them and said, "You are way too good to be working for the federal government on this. They cannot appreciate you. Come to the right side and eventually Michelle got there." Silberman Law is an outgrowth of a number of leading attorneys in the EEO space. We all work at national law firms for our careers. And then six years ago, we started Silberman Law with the idea that we can do this for employers in an innovative and nimble and creative and frankly, more cost-effective way as a boutique firm rather than being part of the big machine of big law.

(03:35):

So Silberman Law focuses on all areas that we call the law of the workplace. Anything that touches on workplace law that we focus on, but in particular, what we think of as EEO and systemic EEO. Where are the barriers to equal employment opportunity in your employee lifecycle, whether it be hiring or recruitment or AI tools. And we have a number of law firm partners, lawyers who have done this for their career, but we're also unique in that we've brought statisticians in- house and we're a one-stop shop for all the EEO lifecycle needs, both in terms of defending, but also complying, internal analyses, assessments to see where you are. Are there any compliance gaps and helping employers get that right?

Jeff Gill (04:26):

Fantastic. And I know a lot of DirectEmployers Members rely on Silberman Law for your expertise and really the level of service that you provide to them. They've learned over the years that they can trust Silberman Law and turn to you for expert guidance. I know 2025, I've seen so many legislative changes and a lot of Members are still trying to sort out what does it mean? For our listeners, can you give us a high level summary of what changes we've experienced and the most significant regulatory shifts that federal contractors have experienced this last year?

Mickey Silberman (04:59):

Sure. I'll take the lead on this first one. So it's a good question, Jeff, but I want to add one more piece to it. Most employers, when they ask the question of what happened in 2025? That's good. That lays a foundation. That is just, to me, the beginning. Or let's say late last year we got to the end of just the beginning. 2026 is actually going to be where the rubber meets the road because with every new presidential administration, they come in, they assess what the previous administration did, and it takes time. It's a big ship to change the direction of. So what are the highlights? We know the most obvious one. Executive order 11246 was rescinded on Trump's first day in office. Affirmative action for race and gender went away. What didn't go away? Affirmative action for veterans and individuals with disabilities that very much is still embedded in the law.

(05:58):

I occasionally speak with an employer, sometimes in- house counsel who says affirmative action went away. Well, no, part of affirmative action went away, but another part is critical to ensure compliance. And we'll talk a little bit more about that, but it's important to understand what is there and what is not. But here's the part that 2026 becomes really interesting. The Trump administration was trying to get the right people in place at OFCCP, at EOC, at the Department of Justice. Those are three primary federal agencies that we're going to deal with that some of your Members are going to deal with this year when it comes to audits, investigations, potentially litigation. And the big picture takeaway here is while some affirmative action elements have gone away and some remain, DEI programs have been challenged, but some still remain with employers. We're really going to see 2026 as the year where the Trump administration executes on its vision for audits and investigations and other challenges for employers.

Jeff Gill (07:07):

That really lays it out for us. And Michelle, I know you've been focused on looking at 2026 and trying to figure out what federal contractors should be concerned about. Can you pick up where Mickey handed off and give us your comments on 2026?

Michelle Duncan (07:20):

Yeah, absolutely. So for 2026, I think we need to ground ourselves in some of the basics. There's a lot of noise. There's a lot coming at us in terms of potentially new requirements, new investigations, new initiatives under the Trump administration, uncertainty around the future or activities of OFCCP and EEOC. But I think this is really a time to not be blinded by all of that and instead really think very pragmatically. Let's be pragmatic in 2026. Let's think about the lowest common denominator. Look, we are all professionals in this space. We know what non-discrimination looks like. We know how to read and interpret regulations. We have expertise in that area that goes far beyond what many give us credit for from the Trump administration. And so what I would say is let's understand what the basics of anti-discrimination means, what the basics of affirmative action means, and let's make sure that we are interpreting that correctly and being really pragmatic when we do it.

(08:33):

So maybe block out some of the noise and the rhetoric that we get from some of the headlines that we read and instead really think about what our obligations are and how we can navigate those obligations as the good compliant citizen of this community.

Jeff Gill (08:52):

That's a great way to put it. I know the Members are looking for a practical way to administer and prepare for what's coming in 2026 as well. So that really hits home. Are there specific actions they should be taking in order to prepare for 2026?

Michelle Duncan (09:05):

Well, my favorite thing to do is to look at what the administration is saying and then see where we can pull on those threads to our benefit. And so I'll give you one really good example, Jeff. This administration has doubled down on merit-based hiring.

Jeff Gill (09:23):

Yes.

Michelle Duncan (09:25):

Mickey, I can't think of a single employer that we work with that would not agree that merit-based hiring is something that they can get behind. And so what I would do is pick up on that thread, pull on it, make sure that your communications, your outward-facing messaging, your training for folks who do hiring and recruitment all align with merit-based hiring. So I mean, I think there are some really kind of neat opportunities here to be strategic and to ensure that an employer's stated objectives align with what the administration cares about. I would say merit-based hiring is something I'm talking to a lot of employers about because that often relates to the use of pre-employment assessments, like skill-based assessments. And while those can be wonderful, there are some obligations and some vetting due diligence work that comes along with making sure that you are using an assessment or a test that was properly developed and is legally defensible.

(10:34):

So that's one of my areas of expertise. And that's how I got into the AI space because that framework that applies under Title VII of the Civil Rights Act to selection procedures equally applies to the use of AI tools. And so that framework that I'm very comfortable with and very familiar with from actually bringing cases against federal contractors and also over the last 14 years, defending and providing advice and counsel to employers who use pre-employment assessments is all coming to bear as we talk about the use of AI. And so this is one of those situations, Jeff, where it's like, let's take inventory of the tools that we actually have in our own toolbox and make sure that we're using those in this new era where you may be talking more about topics like AI that some folks may not be as comfortable with. So I would say merit-based hiring and also how do we not only use AI because a recent SHRM survey showed that the majority of employers are using AI and they've been using it for years.

(11:45):

The question that I get as an employment lawyer is, "Okay, we know the rest of the organization is using AI. How do we use it in HR and how do we use it well?" And I think we're going to talk about that later as we do a deeper dive into the AI topic.

Jeff Gill (12:04):

And use it in a compliant manner, right? Because the law doesn't change if the tool changes, right? So that's an excellent point. Mickey, switching gears for just a moment, we understand there's a lot of proposed rule changes affecting Section 503, including potentially eliminating the disability self-identification requirement and the 7% utilization goal. What's the current status of these proposals and what should contractors be doing while these rules are pending?

Mickey Silberman (12:31):

Yeah. So first, let's juxtapose it, Jeff. There are two different sets of rules, one for military veterans and run for individuals with disabilities, and they have very much worked in parallel. They've looked the same. The obligations have been the same for the last 13 years or so since changes, regulatory changes. The Trump administration proposed changes both for VEVRAA, for veterans and for individuals with disabilities. But to your point, for VEVRAA, they basically left it untouched. You would still, if everything goes in exactly the way the Trump administration proposes, your Members are still going to have to solicit veteran status from applicants, from employees. They're going to have to analyze that data. They're going to have to prepare an affirmative action plan that includes those analytics. The difference here is for disability, the Trump administration has proposed doing away with self-ID requirements at the applicant and the employee stage.

(13:33):

I will tell you this. First, the answer to your question is, we have not yet seen where the administration is going to go. That was proposed, but it has to be approved ultimately by the Office of Management and Budget. Hundreds of comments were put in in response to the proposed regulations. And I'm going to tell you that among employers that submitted comments, 94% of them were opposed to the idea of getting rid of self-ID for individuals with disabilities for a lot of different reasons. But I do know that the Trump administration is weighing that opinion. And it's hard to get employers necessarily all to agree on any one thing. They pretty much all agreed and said, "Don't do this. We want that data. We want to be able to analyze it. We want to be able to benchmark against it so we can assess the effectiveness of our program." And I'll leave you with that.

(14:33):

Both in the proposed VEVRAA and the proposed individual with disability regulations, you still will have an obligation to prepare annual affirmative action plans. And as one critical element of that, you will still be required to quote, unquote, assess the effectiveness of your outreach and recruitment efforts. Well, there may be other ways, Michelle, that I'm not thinking of, but it seems fairly foundational that if you're going to try to assess the effectiveness of your outreach and recruitment efforts, you need data to tell that story. So there's that disconnect in the proposed regs. So we'll wait and see. I suspect we're going to hear in the first half of the year where the administration is going with it, and there's a good chance that the administration backs off of the proposal to eliminate the self-IDs because they want you to continue to assess the effectiveness of the program.

Jeff Gill (15:31):

That would be wonderful. That would be the administration being responsive to the employers and what they're interested in. That's fantastic. Mickey, federal contractors sign government contracts and soon be required to certify that compliance, that they have compliance with all the applicable EEO laws. It's created significant concern. What practical guidance can you give us about reviewing existing programs and determining what they should maintain? I think we're kind of pulling on that thread, but very much would be interested in hearing the one, two, threes, the what can we do today approach to being in compliance?

Mickey Silberman (16:04):

Jeff, I'm going to tell you this. Among the topics we're going to discuss today, AI, Michelle's going to lead on in just a little bit, but this one, if your Members are listening in and writing anything down, write this down. What do I do in anticipation of certification? And I'm going to give you ABCs or one, two, threes in a moment, and what do I do in response to this very quickly growing opportunity or challenge regarding AI? Michelle mentioned, let's make this the year of being pragmatic. I'll give you a parallel. Let's make this the year of being proactive. Many employers last year said, Executive Order 11246 goes away. We don't formally have to prepare AAPs for race and gender. We don't have to conduct these analyses. However, what the administration has done, and it's really, I think, actually quite clever, is they've said, "Look, let's say there's a DE Member with 50,000 employees.

(17:01):

Let's say there's a DE Member with 2,000 employees. We're not going to impose a one size fits all regulatory mandate on you. So we're going to get rid of 11246 and its regulations, but here's what we are going to do. Any government contractor, and this is a critical point, Jeff. A lot of people were talking about the certification. We're awaiting a certification requirement this year. From what I've heard from the administration, they're intent on doing that. But here's something even more basic. All of your Members who are government contractors sign government contracts. They have existing government contracts and one critical term, one provision or paragraph in your government contracts says you're going to comply with all applicable employment laws and other laws. The administration has taken the position. They don't need a new certification from you to hold these laws against government contractors. In fact, in the second half of 2025, the Trump administration launched what was called a civil rights fraud initiative.

(18:11):

Civil rights being EEO and discrimination law. Fraud initiative focuses on this. You government contractor before you ever get a certification. You signed a contract two years ago. You said that you would not discriminate against your employees, but we see a DEI program that you either have now or you had before where it appears that you were favoring certain race and gender groups in the DEI program over arguably other groups that were left behind. You can't accurately sign a contract that you're going to be in compliance with your EO obligations. This is the administration's interpretation. This is important. You can't sign that contract and be accurate in it if you are engaged in DEI or other efforts that potentially discriminate against groups that were left behind. That's a big part of the government's initiative. So here's the bottom line. Your Members right now have current government contracts.

(19:12):

Every one of them is subject to an investigation by the government under this civil rights fraud initiative. We know of about a dozen employers that already are undergoing these deep dive investigations, but as I mentioned, 2026 is going to be the year of execution for the administration. We're going to see a lot more, not just planning, but actual effort and action by them. So what can employers do now? Those employers that said, "We're just not going to do any proactive analysis anymore." We're not going to look at our lifecycle and say, "Is there a barrier to black applicants or white applicants based on an AI tool or a pre-employment assessment or a structured interview that we use in the hiring process?" The bottom line is what I'm telling ... And by the way, the great majority of our clients are doing this now. Under privilege, conduct proactive analyses to see where there are any EEO compliance gaps and then fix those gaps before you sign your next government contract, before you sign a certification, because that will position you to be able to demonstrate to the government that your certification or your contract signature accurately reflects recurrent compliance.

Jeff Gill (20:32):

There's a lot there to unpack. There's a lot there to unpack. And as you said, with President Trump signing the executive order on December 11th for ensuring a national policy framework for artificial intelligence, it gives another layer, another of compliance that we need to pay attention to. So Michelle, I think you've done quite a bit of analysis on reviewing the legislation around AI and how that specifically fits into the states that are creating AI laws and how employers who are multi-state employers can balance their obligations to the state and the federal government and what that means for federal contractors operating in multiple states. Can you expand on some of that?

Michelle Duncan (21:10):

Yeah, absolutely. So there's interesting backstory on this. So I have to start there. So this executive order that was issued by President Trump, it came after two failed attempts for Congress to put into place a moratorium on state's ability to regulate AI. You all may reMember under the big beautiful bill, the one big beautiful bill, the Senate voted, I believe it was 99 to one, overwhelmingly against a moratorium on the state's ability to regulate AI. And then it came up again in a budget, some budget legislation. I think it was the national ... Shoot, Mickey, do you reMember what the name of the bill is? It might have been even the defense spending bill.

Mickey Silberman (22:03):

I think it was the Defense Authorization Act. Yeah.

Michelle Duncan (22:06):

That's what it is. And so twice this was voted down in Congress. And then Trump came out in December and issued an executive order that essentially did what Congress was not able to do. And really the impetus behind this is that many states and some localities like New York City have passed laws or regulations, amended regulations like in California that have created new compliance obligations related to the use of AI. Going back to this, and this isn't new, some of these started in 2019, 2020, right? But there are some really aggressive ones that are about to go into place in 2026. Actually, right in ... So Mickey and I are in Denver, Colorado, and our own state passed what is really the most comprehensive law related to the regulation of the use of AI. And it goes into place later this year. And so, and I'll tell you, what we observed in Colorado was the biggest concern was not the concern, Jeff, about what employers need to do in terms of making sure that their tools don't run afoul of anti-discrimination laws.

(23:31):

That wasn't the focus. What the focus was in Colorado was the tech industry came in and lobbied very hard to ensure that there are no new obligations that extend to them as developers of these tools. Okay, so that's the backdrop. So what the executive order does is the executive order says that in an effort to protect our country's ability to innovate in the AI space, it wants to create a national framework to centralize the regulation of AI at the national level and to prohibit or prevent states from creating a patchwork of obligations that would apply, that will be varied

(24:21):

And will be difficult for folks to navigate. But I do want to just reinforce here that I think that the president's concern here is not about an employer's use of AI and sort of whether that would wreak havoc on innovation. Instead, it really, and I'm reading between the lines here, but this is my take, is that the concern is that if we require much of vendors or developers of these tools, that that is going to essentially shackle their ability to innovate in a space where we're in a global race to be the best at AI. So the executive order itself doesn't really lay out any kind of clear obligations for employers or users of AI or even vendors or developers of AI, but it instead creates a set of instructions for executive branch agencies. And there's several that will have a role. DOJ will have an AI litigation task force.

(25:24):

The Department of Commerce is meant to kind of take inventory of all of the various state and local laws that exist. The FCC and FTC are meant to provide guidance on whether there is a way to preempt, to make an argument that those state and local laws are preempted by current federal legislation. And there's also sort of a promise, if you will, that the administration will seek to pass in Congress a uniformed federal AI legislation. Anyone who has followed at all like the debates about state's rights versus the federal government, you have a red flag going off. And we expect that there will be some challenges faced by the Trump administration related to that very issue. Can the federal government by executive order preempt state and local law when there isn't yet a current uniform AI set of rules on the federal level? And so what I would say is this, we have dealt with patchwork of laws in the past.

(26:33):

Think about ban the box. Think about pay transparency. Think about some of these other issues that seem to kind of bubble up at the state and local level in a different way than on the federal level. And my advice to employers is the same now as it was when I advised on those issues. And it is, you have to understand what laws apply to you and you have to comply with the laws that are currently in effect, right? So if you are in a state or if you are in New York City and you use an automated employment decision tool that's covered by that local law 144, you need to do what that law requires you to do unless and until we get some clear guidance otherwise. So unfortunately for the short term, there's going to be sort of the status quo. I do think, and I'm going to be kind of Pollyanna here, right?

(27:26):

There may be something good that comes out of this. I mean, ultimately if we did, and this is a big if, but if the Trump administration were to encourage Congress to pass, and that is almost like an insurmountable objective. But if Congress were to pass legislation on the federal level that clearly outlined obligations related to the use of AI for both vendors, developers, as well as users of AI, that could reduce this kind of regulatory sort of fragmentation that we're facing. Otherwise, if that doesn't happen, what we're going to need to do is make sure that we are both complying with the state and local laws, as well as ... And here's the biggest takeaway for DirectEmployers Members. Your biggest risk, in my view, and based on litigation that has been brought to date, is under federal anti-discrimination statutes. I'm talking about Title VII of the Civil Rights Act.

(28:31):

I'm talking about the Americans with Disabilities Act, the Age Discrimination and Employment Act. These are statutes that have been around for decades that we know how to comply with. And the courts have said that those laws are so broadly written that the mechanism used to discriminate, whether it is just a decision by a person, whether it's a test, whether it's an AI tool, sort of doesn't matter. So they are written broadly enough to cover AI tools. And the reality is there is litigation that is ongoing as we speak under those three statutes. So what I would say is certainly we have to understand what laws apply and what we're required to do under those laws. But the lowest common denominator here is you need to make sure that your AI tools are not going to trigger complaints under decades old civil rights laws. We want to also make sure, and Mickey may get to this later when we start to talk about EOC, but we want to make sure that we're looking at this in a very broad way.

(29:38):

Gone are the days. I hope none of you ever fell into this thinking because I've always believed it was inappropriate, but gone are the days where we can prioritize certain groups or certain violations over others. For example, we have a tool, but it tends to screen out men or tends to screen out white applicants more so than women or minorities. So you The risk is lower so we can continue to use it without doing our due diligence. Those days are gone. And so we really need to be thinking through what tools we have, the impact of those tools on all groups, as well as whether the tool is actually bringing a benefit to the employer. Is it effective? And then also using that framework that we have under Title VII and under the uniform guidelines on employee selection procedures that outlines how employers can mitigate risk and have legal defensibility related to AI tools as well as other selection procedures.

(30:39):

So that's the long and short of it, Jeff. I mean, it's complicated, but we are going to continue to watch this. There are some deadlines in the executive order for those executive branch agencies and we will just sort of wait and see how this plays out. But for employers, what I would say is you should not wait and see. You should have a strategic, proactive plan for how to comply with the laws that we know are in effect, both at the state, local level, but most importantly at the federal level.

Jeff Gill (31:15):

I like the way you lay that out. And as Mickey said earlier, it's about being proactive this year and understanding the federal and state obligations that employers have regardless of where they're doing business, being proactive with AI implementations to make sure that the AI tools are being set up in a way that comply and are tested to comply with federal and state obligations. I'm just summarizing what you said, but I think that's the gist of it. That just makes good sense. So I think that's excellent.

Mickey Silberman (31:42):

Jeff, I want to add one thing on that because Michelle and I do presentations for different HR type or employment groups. Your Members, the folks that we've interacted with tend to be on the talent acquisition side of this. So the issue of how do we recruit? How do we consider applicants? How do we select them? That area is particularly ripe for litigation. To Michelle's point, we'll see what the Trump administration continues to do in this area. We'll see what Congress does. We'll see how the states react. All of those, there's some element of unknown there, but I will tell you what's known. What's known is that plaintiff's attorneys are looking at this issue very closely and Michelle and I expect to see a significant increase in private litigation against employers under those existing non-discrimination laws for these new and novel AI tools. And Michelle, while we're looking at AI tools in the performance management part of the life cycle, in the advancement, in the training part, the place where there is the greatest exposure, clearly, because the numbers are big, is in the area of recruitment to applicant, to hire.

(33:10):

And that sits very squarely where your Members and their expertise in talent acquisition sits, how their organizations are relying on them. So I can't emphasize this enough. Anytime that a TA team is considering an AI tool in those first couple of steps in the employment process, pull in your counsel, put it under privilege, consider what this tool may do in terms of creating exposure in addition to will it help you achieve your recruiting and business goals?

Jeff Gill (33:45):

Right. That's well said. Recently had a conversation with an organization who was implementing a talent acquisition AI tool and the executive on the employer side asked the AI provider, what's the efficacy of the decision making percent? So there was a pause.

Michelle Duncan (34:07):

Yeah. I mean, Jeff, what you're describing is something that we see far too regularly. There was really a rush to the marketplace on a lot of AI recruiting tools. The need was identified and then it feels like everyone and their brother had some tool that they were going to bring to market and some of them are really good. Some of them are not. I mean, you literally would have a difficult time explaining how the tool works. You would have a very difficult time making out a case that the features that are evaluated by the tool are job related, which is such a basic, basic foundational element of any kind of selection procedure that you might use. But more important than that, we also have pushed back. I mean, I spend a lot of time with employers and vendors on Teams calls, on phone calls, where I'm asking questions, just doing kind of a vetting process for the employer.

(35:13):

And more often than not, what we find is that the vendors are not able to prove up that the tool is actually effective. And so part of what we offer to employers is that efficacy piece, right? We want to understand, of course, we're employment lawyers, we're concerned about violations of anti-discrimination laws. And so we want to understand the impact of the tool, and that would be what I would call a bias audit. And that's part and parcel of what we do. But we also look at the efficacy of a tool. If it is not making or assisting a human decision maker make a decision in a way that's going to benefit the employer, why would you use a tool when it's going to put you in the crosshairs of all of these local, state and federal priorities for enforcement? So just making sure that it is both effective and also non-discriminatory is what we're looking at.

Jeff Gill (36:23):

Yeah, that's a great point, Michelle. Great point. So clearly if I'm a DirectEmployers Member and I'm looking to implement an AI tool, reaching out to council, reaching out to Silberman Law to say, "Help me through this process so that I choose the right tool so that I put it in place correctly." That's a big part of a program that I should include in my program. What other resources can I use or publications can I use to educate myself, to monitor the program, to really get a successful implementation that's not going to put my business at risk because the AI tool isn't doing what I would like it to do?

Michelle Duncan (36:56):

Sure. So one resource, and I'm going to warn you, it's kind of a beast, but it is a really great document. Mickey and I are both part of, Mickey's actually one of the co-founders of the Institute for Workplace Equality, which is a Member organization that does a lot around educating federal contractors on their obligations and what have you. They actually put together a task force on the use of AI. And this multidisciplinary committee included, I think there were one or two plaintiff's attorneys, a couple of defense attorneys like myself, government officials, biopsychologists, statisticians, experts in machine learning and data science. I mean, it was a powerhouse committee and we put together and published an AI tac report, which we're happy to share. It is a beast in that it is very comprehensive, but I will tell you, if anyone is looking for guidance around the use of AI in making employment decisions, it's a great resource.

(38:05):

And it's one that we know the government has paid attention to as well. And on that point, I just want to share. Earlier I said we kind of have to block out some of the noise and some of the headlines related to things like the executive order to prohibit states and localities from regulating AI. I don't want anyone to walk away from this session thinking that the Trump administration doesn't care about regulating AI because sometimes you have to kind of peek behind the curtain to see what folks are really doing, right? And if you look at the guidance that has been issued under this administration by the Office of Management and Budget, OMB, Russell Thot is the person who is the head of OMB. He issued a really comprehensive memo to all of the executive branch agencies around what they should be doing if they're using AI.

(39:03):

And many of the things that we are recommending that employers do, like analyzing for whether there is any unintended consequence of disparities based on civil rights categories, race, gender, ethnicity, right? Those very same recommendations have actually been adopted and provided by the Trump administration to the executive branch agencies as they think about how they're going to handle governance of AI for the executive branch. So I think there's a lot of kind of reading between the lines that has to be done here, but there is a really clear framework that can be applied. And it really starts with, one, just inventorying your tools, making sure you know what's being used and you understand how they're being used. Like Mickey said, hiring is really the highest risk, but sometimes there will be a tool that was purchased for coaching or performance improvement. And then we find out that it's actually being used to make pay decisions or being used to make promotional decisions, right?

(40:16):

So understanding and making very clear what the use case is for each of the tools, and then kind of lining that up with what are the laws that apply and what do we have to do under those laws. And then finally, the two other analyses, Jeff, that we were talking about earlier, efficacy, as well as just ensuring that the tool doesn't have the unintended consequence of screening out folks based on a protected category under civil rights laws.

Jeff Gill (40:48):

That's fantastic. So I wanted to recap and talk about what are the three categories that we should do. But Michelle, I think you hit the nail on the head, right? For Members of DirectEmployers who are interested in rolling out tools, I think the summary is to be proactive, make sure they're doing their homework, doing their due diligence on the tool and making sure that it's rolled out properly, and reaching out to experts like Silberman Law to make sure that they understand all the implications of rolling out a tool and managing a tool of that nature. So all in the context of being compliant with federal and state and local laws around hiring. So did I get it? Is that an accurate summary?

Michelle Duncan (41:23):

Oh yeah, I think so. I mean, and even beyond the kind of pure compliance box, and that's important, and there are high penalties and reputational harm that can come from not being in compliance. There is also just a desire to do the right thing, use tools that are effective, but also that aren't going to result in discrimination. And so I think that's what is so incredibly important about the work that we do.

Jeff Gill (41:55):

I completely agree. So one of the reasons I took this job is because it really is about doing the right thing, helping people find jobs, helping find employers find good talent and manage their business. So Mickey and Michelle, thank you so much for your expertise and your perspective with us today, your insight as to what's to come in 2026. I know you both have been staples in the community for many, many years, providing solid legal guidance to so many of our Members. If our listeners want to get in touch, what's the best way to do that?

Mickey Silberman (42:23):

You'll have our contact information and DirectEmployers certainly can share that with the Members and those who join for this podcast. We have a website, go to Silbermanlawpc.com. And we also very regularly put out, we have what's called our employer alert. Many of your Members are already registered for our employer alert system. This is for both clients of our firm, but those in the employer community generally. So please, if you want, jump on our website or send us an email and we'd love to add you for employer alert to keep you updated on the cutting edge changes this year.

Michelle Duncan (43:02):

I've never been to your DEAMcon.

Mickey Silberman (43:04):

Yes.

Michelle Duncan (43:05):

And I will be speaking on AI this year. And Mickey, you'll be there too, right? Yep. So we're going to be doing a couple presentations there and really look forward to spending some time in person with some of your Members.

Jeff Gill (43:19):

And I am looking forward to that as well. So thank you both for joining us today. We really appreciate it.

Mickey Silberman (43:23):

Thanks so much. Thanks.

DirectEmployers Association (43:26):

Thank you for tuning in for another episode of the DE Talk podcast. Stay connected with DirectEmployers on Facebook, Twitter, and Linkedin, or subscribe to receive updates straight to your inbox by visiting DirectEmployers.org/Subscribe where you’ll receive notifications of new episodes, webinars, events, and more.