Torin Ellis on Bias and the Machine

[00:00:00] You know, back in the day, I had a business, six full-time employees, and though I loved the business and I loved the employees, I hated all the paperwork that went with it. I grew up in a Fortune 500 company, and I wanted everything to be that buttoned down, and I could never do it. I was always struggling to keep up with time tracking, benefits, performance management. If you're in HR, you know exactly what I mean. So I can tell you, if you're spending more

[00:00:29] time managing your HR paperwork than your people, now is the time for Bamboo HR. It's the HR tool that streamlines your process so you can focus on your team. Bamboo HR simplifies everything, so you can stop spending hours and hours on payroll, time tracking, benefits, performance, creating reports, and all that. With Bamboo HR, those hours are shaved down to minutes, so you'll get your time back to do

[00:00:57] what you really love. So, are you ready to see what Bamboo HR can do for you? Get a free demo at BambooHR.com slash free demo. Get back to doing what you actually love in HR. Go to BambooHR.com slash free demo. That's BambooHR.com slash free demo. Feeling kind of left out at work on Monday

[00:01:22] morning? Check out the barf. Breaking news, acquisitions, research, and funding. It's a look back at the week that was, so you can prepare for the week that is. Subscribe on your favorite podcast app. Welcome to People Tech, the podcast of WorkforceAI.news. I'm Mark Pfeffer.

[00:01:52] Torrin Ellis is my guest today. He's a noted diversity strategist and analyst in the worlds of HR and HR tech. Plum recently sponsored his research report exploring the role of AI in enhancing DEIB efforts. The report looks at AI's role in bias, how it can help and how it can hurt, as well as the roles of both AI and people in efforts to combat it. What's HR's role? We'll look at that too on this

[00:02:20] edition of People Tech. Torrin, it's always good to see you. To jump in and tell me about this report that you've published. Yeah, again, Mark, the title of the report, the work that we put together is Reducing Bias in HR Using Artificial Intelligence. And what we knew is that

[00:02:44] AI was an extremely hot conversation last year and that the escalation of that pace, that desire, that appetite to discuss such would only increase in the days, weeks, months, and years to come. And so what I decided to do as my first analytical or research paper is I wanted to take

[00:03:06] a positive spin on artificial intelligence. And let me couch that by saying, I'm not ignoring the ethical concerns, the representational concerns. I'm not ignoring some of the ills that are present in the AI solution in general, not a specific solution or technology, but the space in general.

[00:03:31] I'm not ignoring some of the concerns that people have. But what I did when we decided to write the report is we wanted to take a positive spin because we knew that if in fact HR, which is critical to an organization, we needed to get that central function on board sooner rather than later.

[00:03:55] And one of the things that we found in the report is that people in HR are still hesitant. They are reticent to use and to deploy the technologies. And so we can talk about that a bit more, but it was really sort of an urge from a different point of view, my point of view, an urge of them to say, you know what, I want to learn more about how we can deploy these technologies, some of these solutions

[00:04:23] in various ways that we perform in our function. Now, in terms of bias, which is, you know, an age old challenge, you know, in the HR and the workforce world. Can you tell me about AI's role in combating bias? What does it have the potential to do? So, I mean, it has the potential of doing almost anything that our mind's eye, as they would say,

[00:04:52] can envision. And so what we learned in the report is that if you are deploying AI solutions, that you are seeing efficacy and efficiency in how you are curating some of your diversity slates or your interview slates. We're seeing retention change when AI is being applied. We're seeing communication cadence and schedule shifting so that you find less individuals in black holes or wondering where they

[00:05:21] are in the process. We see organizations being able to curate specific or more personalized learning appetites and diets, if you will. You know, I'm hungry to learn this, but my manager doesn't even necessarily know that I need to know this, that I should know this. And yet when we're looking at the data on the backside, AI solutions are surfacing. Here's potential paths of learning that an individual

[00:05:49] might go down. We're seeing AI help with scheduling. You know, there may be some demands on a Mark Pfeffer because Mark is taking care of an aging parent. And Mark may not be able to attend meetings on Monday mornings or Friday afternoon happy hours or certain activities on a Wednesday. The AI solution is looking at that. So what we are seeing is a more personalized experience. And we're not coddling

[00:06:15] or handholding individuals. But what we are seeing is that the AI solutions, when deployed right, it can add all types of engagement. It can increase levels of productivity. And it really can situate people to say, I'm going to stay with this organization longer because it appears from everything that I'm looking at, it appears that they really care about me. So we do see AI as having a great deal of promise

[00:06:42] within an organization. Okay. But can you tie that back to bias for me? How can AI, with all those dynamics said, how can AI impact bias that's taking place in an organization? It's the exact opposite. So if you run the replay of every single thing that I said, imagine not being chosen for special

[00:07:07] assignments. Imagine being overlooked for learning and development opportunities. Imagine having to really focus on your parent. And unfortunately, your manager, your team doesn't really recognize that you need a bit of a modification in your schedule. It's not that you're disconnected. There is a life event that's happening. It's a real time life event. And so the AI can really help unearth and maybe in some

[00:07:35] cases, hopefully in more cases, Mark, it can avert the bias that humans may display in the scenario. I mentioned the communication cadence. There are some people who may not receive communication because of their zip code or because of the school they went to, or because of the year they graduated. The AI can thwart some of that human interference, if you will, interrupt that bias and make sure that

[00:08:03] individuals are treated equitably across the entire franchise and organization. Does that make sense? It does make sense. Though it does lead, I think, to my next question, which is about the AI itself and the bias of the AI itself. You see reports where biases crept into AI doing one thing or another related to employees or DEIB, how does that concern fit into this whole discussion?

[00:08:32] Well, I mean, it's a fair concern. And again, when we talked about why we did the report at the very, very top, we're not minimizing that representation is extremely important. We're not minimizing that ethical AI is something that we should absolutely be focused on inside of our organizations. We know that if you do it the right way, ethically, that you'll have AI solutions that are technically

[00:08:56] proficient. You'll have solutions that are beneficial to not only the organization, but beneficial to society and most importantly, beneficial to the individuals that are being impacted. And so what we know is that we think that organizations should go through a deeper and more involved discovery.

[00:09:17] They should not just rely on the sales pitch of the organization moving the solution, that they should be in a position where they are asking some very detailed and involved questions around data sets and cleanliness of data and refresh of data. And who was involved in the development of the solution? Who's going to be involved in the deploying of the solution

[00:09:43] internally? How are we going to evaluate the efficacy of that solution? What are the metrics to say that we are making progress? How do we look at the ROI? How do we recalibrate? And so I am absolutely not a person who is closing my eye to the presence of bias, to the presence of infractions that are disingenuous, to the harm of AI possibly being perpetuated.

[00:10:10] But I am also not a person who's going to stand flat-footed and handcuffed, paralyzed, and not willing to make a decision. I feel like we can do both. We can be critical of the solutions that are out there. We could also be discerning in terms of how we make our decisions. Now, as you're looking at all of this, who should be guiding these efforts, these discussions? Should it be the AI or should it be a human of some sort?

[00:10:38] No, I think guiding should be done definitely by humans, but I think it should be a cadre of humans. I don't think that it should only be a chief technology officer. Listen, we know from our research, our research lines up with Gardner's research. And for the last several years, technology challenges or awareness or adoption has been an Achilles heel in HR in the top five for the last several years.

[00:11:08] So our research lines up with that of Gardner. And so I don't believe that because HR folks may not be as technically astute, that they should be left from the table. I believe that the people at the table guiding these AI solutions, the implementations, the spend in the first place, but the implementations, the measurement, the recalibration, and the go-forward strategy, I believe it should be a cadre of people sitting at the table.

[00:11:38] Sure, someone from HR and definitely someone from legal and people that are going to be impacted by the solutions themselves. I believe that we should have some outside consultants sitting at the table who have no relationship to the organization, but they only want to see the good of the organization, the success of the organization. So bring in some non-bias or non-partisan consultants that can complement your AI ethics or your AI advisory team, if you will.

[00:12:07] If you like swiping, then head over to Substack and search up Work Defined, WRK Defined, and subscribe to the weekly newsletter. Looking for the inside scoop on payroll? Whether you're a payroll pro or just curious about your paycheck, we've got two podcasts you cannot miss. It's about payroll delivers the latest trends for payroll pros, and it's about your paycheck is here for employees. Breaking down pay and rights in simple terms.

[00:12:34] Two great shows, two great hosts, and endless insights to keep you informed and empowered. Subscribe now and elevate your payroll knowledge. Brought to you by Work Defined, where payroll meets people. Can you give me a little bit more detail on what you think HR's role in all of this should be?

[00:13:02] You're talking about teams, basically. A lot of people around the table working on these efforts. How's HR fit in? So the first thing that I would say is stakeholder, stakeholder, stakeholder. That's the very first thing, and I'll repeat that. Stakeholder, stakeholder, stakeholder. Mark, what I want to see as we escalate our use of AI and the promulgation of these various solutions,

[00:13:29] what I want to see is an HR that returns to its roots of protecting stakeholder. Some individuals, that can be a bit of a challenge because we say to ourselves, well, we got to protect the organization. Maybe, but I think you protect the organization by primarily protecting the people. And I don't want to see issues of another Me Too movement. I don't want to see what we were seeing in the beginning days of COVID,

[00:13:58] where microaggressions were going up against women in Zoom environments. Zoom environments feeling even more accosted, even more aggrieved by their counterparts. How does that happen? And so I do want to see HR play a very strong and demonstrative role in saying we are going to explore how these AI solutions impact the various functionality of our discipline.

[00:14:26] And we are going to make sure that these solutions are deployed in a way, number one, that takes care of our stakeholders. And number two, takes care of our shareholders and organization. And I believe if HR shows up in that particular way, place any flag in the ground, Mark, that the people are important. I believe that our organizations, all of them, will become better.

[00:14:55] One of the things that I have observed, I guess, AI is very hot right now. And sometimes just saying AI is the answer. In these kinds of discussions, doesn't that mean AI can be kind of a crutch that somebody just says, well, you know, we've got AI involved and AI is going to handle this for us. So they can sort of duck doing a lot of the real work themselves. I mean, you know, of course that can happen.

[00:15:25] But Mark, that can happen with almost anything. You know, if individuals feel like there's a way, a quicker way for us to do it, if there's a shortcut, let's make it happen. But I also want to address that from a different standpoint. Inside of our report, we reveal to readers, we used chat GPT for about 12 to 15% of the 7500 plus words.

[00:15:52] We don't want people to feel like they can't reveal and be transparent about how they are leveraging these various technologies. I don't want people to appear like they are smart, that they are not relying on the support of some of these technologies. I just don't want it to become a crutch. Not at all. I'll give you an example.

[00:16:16] Inside of the report, Plum shares with us one of their global financial services firms, 90,000 employees represented in 30 different countries. They were struggling, Mark. They were struggling to get representation in their pipeline. They were struggling for a variety of reasons. For the sake of time, we won't go into that. They reached out to Plum and they just simply said, you know, is there a way that you can help us?

[00:16:42] Is there a way that you can leverage what it is that you are doing so that we are not only increasing the talent that's in our pipeline, but we're expanding the perimeter and where individual applicants and candidates are coming from. And so Plum got into effect. They got into motion. Once they were deployed, the organization saw efficacy and efficiency skyrocket.

[00:17:09] They saw their representation in their pipelines skyrocket. They saw their footprint move from 12 academic institutions to more than 30 academic institutions, all because they did a couple of things. Number one, they revisited the value being placed on resumes. Is there a different way for us to evaluate candidates? Because we know that not everyone is writing the same way, using the same language,

[00:17:37] categorizing the work that they are doing in the very same way. Are we missing some individuals? So first, it was a reprioritization around resumes. And then it was placing in effect a 20-minute assessment using their technology, which is definitely on side of an AI platform. And all the organization saw was an increase and increase and increase. And so I do want people to understand when you deploy it the right way,

[00:18:06] even if you're looking for a shortcut, so to speak, even if you're looking for efficiency, efficacy, if you deploy the solution the right way, you can win, you can get what it is that you want, and you can still remain honest to your mission and the KPIs that you are measured by. So you've done all this research and put together this excellent report. What were the biggest takeaways for you once you were said and done?

[00:18:36] Yeah, I will tell you there were three. I'm going to give you a positive and then two negatives. The biggest positive was that retention. Retention was impacted in organizations that have deployed AI. And I want to be very specific about that. In our research, when we did our query, we know that not everyone has deployed AI.

[00:19:02] But inside of the organizations that answered in the affirmation that they have deployed AI to help reduce bias and increase engagement, in those that responded, yes, they have seen an absolute increase in retention. So that was beautiful for us to know. And certainly there were a number of other categories that had been positively impacted, but the retention was a promising revelation. That's number one.

[00:19:31] I think the two things that disappointed us in the research. Number one, not a number of organizations. Let me say that differently. There were a number of organizations that did not connect their D&I efforts with their AI exploration or decision.

[00:19:56] So they went out to the marketplace, looked at a plethora of tools or solutions. And diversity and inclusion efforts, metrics, KPIs were not a part of the checklist of considerations. And I do believe that if you are going to build a strong organization, an impenetrable organization,

[00:20:24] that once again is measured by engagement, is measured by an increase in productivity, and is measured by positive attrition or retention, then I do believe that at some point in your exploration of new solutions, you should ask that organization, what are they doing to address, to impact, to enhance, to help you with your internal D&I efforts.

[00:20:51] That so many organizations did not have that as one of their checklist considerations was a surprise for me. One of the other surprises for me was still the number of organizations that were on the sideline mark as it relates to deploying AI. And so I feel like we still have a very heavy lift of inviting people to the table, to the learning,

[00:21:19] inviting people to be able to make some decision that may come with a bit of risk. It may be a bad decision. It may turn out to be a less than effective solution. It may be one that really amplifies and highlights bias in our particular organization. Whatever that risk is, whatever that impediment is, whatever that trepidation is,

[00:21:48] I want people to shake that and to make a decision. And we just found that far too many individuals and organizations, Mark, were still sitting on the sideline hesitant to get into the fray. And what I'm afraid of is that those organizations, yes, they'll be able to play in the game this month, or they'll be able to play in the game six months from now or 12 months from now. Everyone doesn't need to be first mover. Everyone doesn't have to jump on the hottest solution.

[00:22:17] But what I want is for organizations to feel like they are in a position where they can make decisions that once again are going to go back to taking care of their stakeholders and find them in a more favorable, I guess, lot or posture as it relates to the marketplace. Warren, thank you so much for talking with me today. It's always great to talk to you. The report is fascinating.

[00:22:45] And if people want to get it, where should they go? So it would be great for them to go to Plum.io. Again, that's Plum, P-L-U-M dot I-O. And if you want to find me, I'm at Torin Ellis across all of social media. And I can promise you over the next, I don't know, several months, we're going to be posting periodically throughout the week. So I'm certain if you follow me on one of the platforms, you'll see a post and it'll lead you right to the report.

[00:23:16] No, it's a fascinating report. And thank you for sending it to me. Thank you for spending the time with me today. And I hope we'll talk soon. Absolutely. And you know, I appreciate you. Thank you. And I wish you all the best. My guest today has been Torin Ellis, a diversity strategist and analyst covering HR and HR tech.

[00:23:46] And this has been People Tech, the podcast of WorkforceAI.news. We're a part of the Work Defined Podcast Network. Find them at www.wrkdefined.com. And to keep up with AI technology and HR, subscribe to Workforce AI today. We're the most trusted source of news in the HR tech industry. Find us at www.workforceai.news.

[00:24:15] I'm Mark Feffer. I get it. The podcast just isn't enough. That's all right. Head over to your favorite social app, search up Work Defined, WRK Defined, and connect with us.