DE Talk | How SMBs Can Use AI Hiring Tech in Inclusive Ways

Season 6 • Bonus Episode 2.5

Businesses of all sizes are impacted by the potential benefits and possible harms of artificial intelligence (AI) in their recruiting and hiring technologies. While some may consider the HR tech space “too technical” or beyond their current capabilities, this mentality can cause them to lose job candidates with the skills they need to succeed. In some cases, small and midsized businesses (SMBs) may already use AI-enabled hiring technologies without knowing it. In this session, Our Ability’s CEO, John Robinson, and the Partnership on Employment & Accessible Technology’s Director of Communications, Lex Huth, team up to show how people with disabilities are leading the way to more inclusive AI in hiring technologies — and how SMBs can maximize the benefits they get from these AI-powered tools.

About DE Talk

For DirectEmployers, it’s all about valuable connections and meaningful conversations. This monthly podcast features honest and open dialogue between powerhouse industry experts on a variety of HR topics ranging from OFCCP compliance advice to emerging recruitment marketing trends, diversity and inclusion initiatives, and insightful solutions that help infuse new life into your HR strategies.

Episode Guests

Episode Transcript

DirectEmployers (00:00):

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

(00:09):

Artificial intelligence is everywhere, and while the headlines often spark fear and uncertainty, the real story is far more exciting. When harnessed thoughtfully, AI has the power to unlock possibilities far beyond human limitations. In this episode recorded during our 2024 annual meeting conference, Lex Huth and John Robinson dive into an eye-opening conversation on how AI can be a game changer for inclusive hiring. Discover how this evolving technology when used with intention can help create a more diverse and equitable workforce.

John Robinson (00:41):

Good morning. Oh, come on now. End of the day or end of the session. Good morning. Here we go. Excellent. It’s good to be here. It’s been a long time. I gave a speech on a stage like this for DirectEmployers years and years ago, so it’s nice to be back and nice to be here. Lex invited me to co-host this. So you’re going to hear from Lex and I over the next little while about how small businesses, small and medium sized businesses can use AI hiring tech and inclusive ways. It’s something that I think about all the time. Quite frankly, we as a company have made a conscious decision to use artificial intelligence to add people with disabilities into the dataset, and that we are people with disabilities that own the company and we are people with disabilities that are coding. And so it’s extremely important to us that we think about this from three sides and hope we convey that with you.

(01:47):

I’ve owned Our Ability for probably 13 years now, and it’s been a wonderful experience and I’ll explain that a little bit with you as we go along here in the next little while. So our session flow here is we’re going to do some introductions and I’ll do mine in a second. We’re going to do a little bit of background on artificial intelligence. We’re going to do hiring risks and how to handle them. We’re going to do inclusive AI for small and mid-sized businesses. We’re going to talk about optimizing your job descriptions. And this is important work that we think about at Our Ability on quite frankly hourly basis. How can we write a better job description so that we can match to our community? And then we’ll take your questions and answers after. I’m John Robinson. I am president and CEO of Our Ability. As I said, I started Our Ability about 13 years ago.

(02:42):

It just came up on the 13th anniversary. I started it ostensibly to be a mentor network. Our goal was to mentor people with disabilities towards employment. I came out of 16 years of media, so producing videos for public broadcasting. And the idea initially was to create video stories of people with disabilities who were successful so that we could mentor a younger generation. Quite frankly, what ended up happening was very transactional. The companies that I would speak at around North America for the better part of 10 years really just asked us, Hey, can we put jobs up on your website? We know that you’ve got traffic on your website. Can we put jobs up on your website? So stupidly, I said, yes, we can do that. And then what also happened on the other side of it, the younger generation of people with disabilities who I thought we would mentor, quite frankly, just email us every day.

(03:42):

And there’s not a day that goes by that I do not get an email in the past 13 years. Can you help me find a job? Can you help me? Can you help me? I don’t know how to navigate my existing services. I just want help finding a job. And so that has held true. And so if there’s one thread that I can pull through the story of what we’re going to talk about here today is that people with disabilities still want to be included, part of an inclusive workforce. And that story has not diminished. And so I’m proud to be here and I’m proud to be here with Lex.

Lex Huth (04:17):

Thank you. You’re welcome. I’m so proud to be, we just met for the first time, even though we’ve been working together for years, so it’s always interesting. But I’m Lex Huth. I am the director of communications at PEAT. That’s the partnership, unemployment and accessible technology. On the screen, I have pink hair. Unfortunately it’s a little bit more boring today. It’s brown. And I’m also holding my Sheba Enu dog. He is my emotional support dog and he is fantastic. I wish he could be here, but unfortunately he’s afraid of clapping. And selfishly, I was hoping there might be some clapping today, so I didn’t bring him. Thank you. That wasn’t a call for that. I didn’t even expect that, but

John Robinson (04:56):

Thank you. No, we’ll bring it in. Let’s do that. Yeah.

Lex Huth (04:59):

So I just wanted to take a moment and thank you all for being here, first of all, for creating this inclusive space. Liz and Liz and Krista have done an amazing job behind the scenes, making sure that things are inclusive here, and they’ve created such a great environment that, plus it’s the one year anniversary of me speaking at DEAMcon for the first time, which I don’t think many people know. Was my first in-person presentation at a conference talking about my disability. It was terrifying, and it still is. I am still scared because of disclosure, because of stigma, because of so many things. But Angie, who hopefully is in this room, and if you can find her, please connect with her. Oh, hey. She came up to me after and she said so many helpful things, but beyond that, she put things into practice over the past year, she came up to me and told me that the things that she was going to do happened, and that is incredible. And so I just want to say thank you to her and thank you to her for inspiring me to keep speaking about this and to keep coming here. So I appreciate it. And now I want to create an inclusive space for all of you. If you need to get up and stretch, if you need to ask a question midway, I might not see you, but get somebody’s attention, Sean will

(06:20):

Let me know. Thank you.

John Robinson (06:21):

I got you.

Lex Huth (06:22):

Get up and stretch. Go get coffee. Whatever helps you show up in this environment in a calm, relaxed way because I know it’s been quite the week and I just want you to be comfortable. So with that, I will get started unless there’s any questions to start with, which I doubt. Alright,

(06:40):

We’re good. So we’re going to start with a bit of a controversial topic, which is the return on investment of disability inclusion. And that is controversial because inclusion should be inherent. It should happen from day one. It’s not something that we should have to ask for, but we all know that shoulds don’t mean anything. We have to show the bottom line often. So I’d like to start with this slide because Accenture has some very interesting insights and reporting on the return on investment of inclusive workplaces. So the quote says specifically companies that have led on key disability inclusion criteria during that time have realized 1.6 times more revenue, 2.6 times more net income and two times more economic profit than other participants in disability ends annual benchmark survey. There’s also a chart that shows those numbers as well, and that just really underscores that disability inclusion does make business sense. It’s not just the right thing to do or the good thing to do. It actually does make sense.

(07:45):

So now let’s talk about why we’re all here. Artificial intelligence, which some people are really tired of hearing about it, talking about it, especially because there’s so many flashy ways that we talk about it. There’s so much intrigue, ChatGPT, all of that. That’s fine, that’s great. But what are the practical ways that we’re using this in our workplaces? So AI has been adopted incredibly quickly. It’s outpaced a lot of the guidelines, the standards, the regulations that are required to make sure that equitable technology design and implementation is used. So it’s just sort of out there now. It’s already in a lot of our workplaces and it hasn’t necessarily met the independent testing criteria that is required to make sure that it passes all of these things, that it is equitable, that it is useful, and that it is being used in inclusive ways. But don’t take this as gloom and doom.

(08:41):

John’s going to talk about some of the awesome things that we can do with AI, but we have to get past these barriers. We have to make sure that we are doing the right thing with the technology instead of just letting it out there and letting it run wild. Because tools that designed with equity in mind can lead to discriminatory outcomes for people with disabilities, but also intersectional identities, other underrepresented groups At PEAT, we’re not just looking at people with disabilities, we’re looking at everybody. We want everybody to be included. And AI can often exclude people if we don’t use it properly. So how are we using AI in our workplaces? Sometimes it’s a little bit hidden, so it can be used often for assistance or assessment. And now is the time when I get to disclose my disabilities again on this stage. So I’m visually impaired.

(09:32):

You’ve probably seen me walking around with a cane, but you’ve probably also seen me seeing I have left central vision left. It’s very complicated. It’s a degenerative disease and that vision is overlaid with my right vision, which is unusable. So I see double permanently. I can read a menu, but I might walk into a trash can is what I say. So it’s very confusing. I’ve just started using my cane and I also have obsessive compulsive disorder or OCD, and both of those things have been both helped and kind of set a little bit back by some AI. So again, really depends on how you’re using the technology. One way it’s been helped is communication tools. So I use speech to text and text to speech constantly. It is a lifeline for me. I thought I was going to have to leave the field because I was training to be a technical editor while I was rapidly losing my eyesight in grad school.

(10:25):

And those things don’t seem to go together, but they can if the technology is built with inclusion in mind. Also collaboration. So we’ve all been in a lot of virtual meetings. There are often AI tools embedded in that collaboration, taking notes, comments, follow-ups, things like that. And also my favorite assistive technologies. So things like seeing AI, computer vision. So I just started taking orientation and mobility classes to use my cane. They were showing me different apps where I can hold them up and see what’s around me. The phone will read it back to me. So seeing AI can be really helpful in personal and workplace settings, and that’s something I’m very excited about. So AI isn’t all gloom and doom. It may have sounded at the beginning. Do you have anything to add, John?

John Robinson (11:11):

Yeah, it’s not all gloom and doom. I understand the fear. Certainly I’m a Battlestar Galactica fan, so if you take that to the nth degree, we’re all going to die because of AI. But

Lex Huth (11:23):

Let’s hope not.

John Robinson (11:24):

Yeah, let’s hope not in the in-between, you know what, it becomes a great equalizer for Lex and I and other people with physical disabilities because the world becomes more mobile. And so the way I like to look at it is, yeah, we’re going to be taken over by Cylon someday, but in the in-between, you know what, we’re going to be able to use assistive technology. We’re going to learn from the assistive technology and we’re going to grow.

Lex Huth (11:46):

Great. All right, so next up, human assessment. So you’ve heard a lot about recruiting and hiring this week and AI is often used in that. John will talk about that more job matching, resume screening, interviews, things like that, making sure that these huge volumes of data are getting read through. But again, there are good and bad ways to do that. There are places where bias can be introduced into the process. As a person who had to take time off because of surgery, I don’t have continuous employment history the way that you might expect, will that weed me out? That is concerning. Also, surveillance is another way that people are using a lot of AI for their workers, especially remote. I’ve heard from some people that they love it. They love having those benchmarks, they love having those metrics, knowing where they are. So it’s not ambiguous.

(12:36):

As somebody with an anxiety disorder, it scares me. It stresses me out. I am constantly worried about it. So you have to know your workers and you have to know what they need. You can’t just assume, oh, one person said surveillance is bad. We can’t use that. You just have to mindfully implement these tools. There was a specific incident where I had a freelance employer who would take screenshots of my screen every 10 minutes I think it was. I didn’t want to use my screen reader that surveillance technology was going to disclose my disability for me. And I was concerned that they wouldn’t want an editor who couldn’t see very well, even though it didn’t change my job performance. So you have to be careful with things like that. And also training. So completing documentation, job simulation, which is one of my favorite applications in vr. If you’re interested in inclusive vr, I just gave a talk at South by Southwest on virtual reality and being able to be in it while being only partially cited and what that means. And I’ve been participating in research on that. So let me know if you want to hear more about that. I love talking about it. Anything else on human assessment?

John Robinson (13:46):

No, we’ll get into it and we’ll get into it.

Lex Huth (13:47):

Sounds good. Alright, so now we’d like to ask you if any of you use AI tools at work and what benefits you’ve noticed

John Robinson (13:56):

On two.

Lex Huth (13:57):

Oh, we have hands. Great.

John Robinson (13:58):

Three.

Lex Huth (14:00):

I was afraid of silence.

John Robinson (14:01):

Four.

Lex Huth (14:03):

Okay, great. Does anybody want to talk about the benefits they’ve noticed

John Robinson (14:10):

Right there?

Lex Huth (14:10):

Okay. Can we have a microphone? Thank you so much. I appreciate it.

Cheryl (14:20):

Hi Cheryl. I disclosed actually for the first time at a conference yesterday that I have dyslexia. Congratulations. And so I didn’t really, throughout school I struggled. I memorized words. I just had no idea what it was. My parents didn’t know what it was. I didn’t find out what dyslexia was until my son was diagnosed with it in first grade. So I use AI all the time in work Now I joke with all of my recruiters and I say, has anyone used ChatGPT today? Or have you used Microsoft Copilot? Because I’m trying to enhance and make sure that they understand that it’s not cheating, it’s just helping me because it takes me longer to write things. I was always the one that was joking when I was writing on a flip chart saying writing on a flip chart doesn’t count if you spell it wrong. Ha ha. Right? Because I knew the letters were going to get reversed for me. So AI has just been really helpful in reducing the amount of time it takes me to get thoughts that are in my head onto paper. But I really, one thing I didn’t share yesterday is that I see it as a superpower. I literally can take information and see it in so many different ways and then be able to pull it together really quickly. And AI does that even faster for me. So it gets me doing more of the human work versus he administrative stuff, board papers that would bog me down and I would probably hate my job. Thank you so much,

John Robinson (15:37):

Cheryl. I saw your presentation yesterday and I was thinking at the time I applauded you not only for disclosing sort of publicly right there on stage right here on stage, but also then explaining how you use AI for its tools for good, right? That’s the point. Listen, we can sit here and think we’re not going to advance. We’re not going to use furthering technology. We’re not going to do this because we’re going to be entrenched in the way we always have done business the way we’re going to continue to do business. But the truth is the next generation isn’t going to do that. They’re going to keep coming along and people with disabilities are going to use it because it becomes an equalizer for us. And that’s a huge element in what we’re discussing. So thank you for sharing yesterday.

Lex Huth (16:21):

Thank you so much. Anyone else?

Angie (16:29):

Hi, Angie from the Chicago Housing on Authority. So with AI, as the young lady mentioned, I use it for job descriptions. And so I didn’t want to at first because I started to feel like she mentioned that I was being lazy. That old school mentality of you have to do the research and you have to do it yourself, or it looks like you’re cheating if you allow someone else or something to do it for you. And so I hired, I had to hire HR assistant, and so my assistant is younger and she started showing me and she’s like, well, why do you do it that way? You can just do it like this. And so just watching her, and she is from the younger generation, and I said, you know what? I relented. I said, I have to start using this so much more efficient. And it’s just so helpful and it allows me now to do things that gets buried on your desk because now I can just use AI to create, and it is still a human element to it. You do have to edit and make sure that it is what you want, but it helps. So I just wanted to share that. Yes, I have now started using AI for my job descriptions.

Lex Huth (17:48):

Thank you.

John Robinson (17:53):

Good.

Lex Huth (17:54):

Alright, so you have a question? Oh, I’m so sorry. Thank you.

Francesca (18:02):

Just a comment. I understand what everyone’s saying about AI, but for me, I guess the jury is still out when it comes to discrimination, knowing that in real life when people have, I was talking about African Americans, children have names that like me, sound like De Sean, Laisha, Laja, nay, all those will AI, the way the current system kicks ’em out, the way the current system discriminates, and even words like disability has AI so advanced that if someone says they have a disability, it doesn’t kick that resume out. Or if they have different names has been shown in like 60 minutes, it kicks it out. That’s

John Robinson (18:47):

Perfect seque.

Lex Huth (18:47):

Yes.

Francesca (18:47):

I still need to further the research.

Lex Huth (18:50):

Thank you.

Francesca (18:51):

And maybe it’s an AI national convention where all heads come together. You just gave us the perfect segue.

John Robinson (18:58):

You literally just teed us up, so

Lex Huth (19:00):

Thank you.

John Robinson (19:01):

Planted question.

Lex Huth (19:02):

Yes. Yeah, it feels like we asked you to do that because yes, it is still concerning. So thank you. Thank you for thinking along those lines because that’s what we need to be doing. We can’t say, these tools are great, these tools can be great, but they can also discriminate. So let’s talk about that. Alright, so risks in hiring. Hiring bias is a real thing. And people with disabilities know that as do everybody, underrepresented groups, it can be impossible to get a job because of things you disclose, things you don’t mean to disclose. Like I said, the gap in continuous employment or the fact that I’ve been working in disability for several years, that who knows what that’s going to do in the algorithm. So unfortunately, the data available on employees is usually flawed. It’s usually like employees. So you’re not going to get a diverse candidate group because these technologies aren’t looking for them.

(19:59):

They’re looking for more of the same. AI also takes shortcuts whenever possible. So it is learning from flawed data and making jumps from flawed data. And so it can quickly become more bias than the training data. So that’s concerning as well. It also reflects the implicit biases of the people who design it. I think we heard earlier in the week somebody saying that if you think you probably have bias, and that is definitely a concern. A lot of people think, oh, well it’s a technology, it’s not a person. How could it be biased? But somebody made it, somebody sat there and made it. So it can have problems, it can have bias. And so we need to watch out for that. And then finally, many vendors also consider the things that they create these tools proprietary. So they won’t let you see what they’ve done, how they’ve trained it, if there were inclusive data sets used. So all of those are kind of mysteries and they’re dangerous mysteries because we need to know if there are inclusive data sets being used, if it’s just going to make the same choices it always has been. And if it’s going to perpetuate that bias that she was talking about. I apologize, I didn’t get your name, but thank you for that, Francesca. Thank you. Yeah, so listen to Francesca.

(21:14):

So did you have anything to add before I

John Robinson (21:17):

No, we’re good.

Lex Huth (21:17):

Okay, great. So PEAT had an AI think tank because we launched a big toolkit on AI a couple of years ago, and we want to make sure that we are always current on what’s going because it’s moving so quickly. As I said, and in this think tank, we brought a bunch of innovators and people with disabilities, AI ethicists together to learn from them and see what they thought about AI in hiring. And here are four takeaways. There’s a much longer report. It’s on the website, which is P-E-A-T-W-O-R-K s.org if you’re looking for it. But first of all, everybody always said, and John can echo this, involve people with disabilities in every stage and people from underrepresented groups. You have to listen. Tiffany said yesterday that if you’re not listening to people with disabilities, they’re not included in the conversation. And that’s a bigger message. It’s inclusion isn’t just inviting them to the table and saying, well, we had them checkbox. It’s actually listening to people with disabilities,

(22:19):

Sharing risk management duties across organizations. It can’t just be the vendors, it can’t just be the employers. It can’t just be the policymakers. It has to be everyone checking, making sure that these things are used in fair ways, that they’re designed in fair ways. We can’t pass it off to somebody else. Also, letting people know when AI is used for good, for whatever it is in your system, how it’s being used, how it’s making decisions, how they can opt out or how they can request a reasonable accommodation. It’s still a hiring process, so you can’t just say, oh, I don’t know how it works. I don’t know if you can have an accommodation. It’s really a requirement. And then finally, staff at all levels need to really develop a culture of AI. Fairness. Again, staying involved, being in the loop, making sure that humans are a part of it.

(23:09):

Anything to add there? Any questions, anyone? I don’t hear anything. And this was a part, John’s group was a part of this too. Our Ability, our disability led innovation report. So we really wanted to drill down and focus on people with disabilities and how they are using AI in their workplaces, what they saw as benefits, as barriers, and we distilled that down into five topics, which really had a lot of overlap with what we saw from the think tank. So that was great and reinforcing. And those five things are ask before you buy again, if the vendor’s not you, that’s probably concerning. It’s something that you need to know that it’s going to be inclusive or as inclusive as possible. We can’t get a hundred percent, always include humans. As I said, you can’t just let the technology run and say, well, that’s good enough.

(23:59):

You really need to make sure that AI isn’t making the final decision for you. It’s just involved in the process. Be transparent with stakeholders. Let everyone know what AI is doing, how it’s doing it, and how they can get accommodations. Focus on accessibility in every aspect of everything you do please. And if you’re not sure how, there are a lot of resources and a lot of disabled leaders with great information out there, LinkedIn is a good source for that. Top voices have a lot to share and they are sharing it. So pay attention to people with disabilities and also address systemic barriers throughout your hiring process, not just in AI. Again, I’m not just saying this and saying, well, that column is everything now go do it. Pete has a lot of resources on it, as do other organizations. We also have a beginner’s checklist on AI.

(24:47):

All of our resources are free. It’s just a front and back. You print it out, you can check the boxes. It’s not comprehensive. It’s a place to start, but we all need a place to start. So please check out that we also have a podcast John was a guest on. We have AI ethicists on just to talk about their lived experiences with AI and where it’s going in the workplace. So I would encourage you to see that. Now, John, I would like to hear from you because that was a lot for me for inclusive AI tips for smp.

John Robinson (25:18):

Take a breath.

Lex Huth (25:18):

I know.

John Robinson (25:19):

All right, we’re good. Francesca, I want to answer your question and I’m going to do it in a story that I’m going to tell. I’m 55 years old. I am a quadruple amputee, as you can see, born without the extension of my arms and my legs. Like 10 years ago when I gave the keynote speech, I was probably bouncing up on stage a little bit better than I physically am right now. I bring my step stool up in here so that I can sit down and relax. My hips hurt a little bit and my body tends to hurt a little bit. But when I was a junior in high school, senior in high school in Manchester, New Hampshire in the mid eighties, that was really my first experience in services, disability services. And my father set up a meeting with vocational rehabilitation in Manchester, New Hampshire. And it’s a federal building in the mid eighties, and it’s every image of every federal building in the 1980s.

(26:18):

You can imagine it’s a basement office with cinder block walls that used to be white but are now brown with one little window up in the corner, sort of this green leather metal chair thing that you’re supposed to jump up into and a wooden desk in front of me and nothing on the desk except a telephone attached to the wall and an ashtray. And so I sit up in the chair and nobody there, and this man comes in the room and introduces himself as Jack Smith. Literally, that was his name. I’ve known three Jack Smiths in my life, and they all changed my life in some way or another.

(26:54):

Jack Smith sat down at his desk, took out a piece of paper and a pencil, lit a Winston cigarette, and asked me my name, asked me what high school I was going to, asked me my grades, never looked up from the paper, asked me what I wanted to do when I grew up. I thought it was a short comment. I said, I’m not growing anymore. This isn’t going to happen. He goes, no, what do you want to do when you grow up? And I said, I really want to be involved in sports. That’s the first moment. Jack looked up from the piece of paper to look at me in the eye and said, is that feasible? No, probably not. He goes, no, wait a minute. He goes, it’s 19 85, 86, whatever it was. How do you watch sports? Do you read the newspaper every day? Do you listen to the radio? Jack Smith asked me a question, what did I want to do? That’s a rare question for people with disabilities.

(27:57):

Almost an impossible question. Truthfully, if I had gone 95 other times out of a hundred to voc rehab, the question would’ve been, have you considered working in a flower shop? Because I know you can run the cash register if you sit on a stool. But Jack didn’t ask that question that day. He asked me what I wanted to do, and he set me on a path to going to Syracuse University to study television, radio, film. That question is the key component to what we’re doing. I mentioned right now Our Ability, we’re 13 years old. We’ve had a jobs board that we’ve worked with DirectEmployers for a long time. The jobs board continues to grow and grow, and we’re honored to have this relationship with DirectEmployers. But five years ago received a phone call from Syracuse University where I ultimately went to school and graduated on alumni board and about to be named a professor of practice at Syracuse for disability, which is kind of nice.

(28:57):

That’s exciting. I didn’t know that. Syracuse called and said, John, what are you working on? We’ve got a capstone project of eight grad students in information technology. We’d like to give you the grad students to work on something real. Wow. Okay. Well, that Jack Smith question has rattled around in my head ever since I started Our Ability, and quite frankly since I started working, what do you want to do? So I said to the capstone students, I said, we’ve got this job board and we’ve got people with disabilities that email us constantly. It seems to me that we could do a better job in asking, what do you want to do?

(29:39):

Had a meeting with IBM. IBM Watson had just one jeopardy to yesterday’s presentation, got a phone call from Microsoft and Microsoft said, Hey, we heard what you want to do through Syracuse. Do you know about the AI for Good program? Would you be interested in applying for the AI for Good program? By the way, we’re going to give you this grant, this gift, and we want to see if you can use machine learning to do a better job in asking that in one question. But more importantly, can you include people with disabilities into the dataset? The most important thing you asked is people with disabilities. People in other marginalized populations have been excluded from data sets. Can you include, well, wait a minute. If I ask the fundamental question, what do you want to do? What does it matter, what my disability is? What does it matter what my first name is?

(30:33):

What does it matter what my heritage is, what my sex is or what my sexual orientation is? It doesn’t. What does matter is there are 50, a hundred thousand jobs out there on our jobs board. There are skills in those jobs, and if we do the right, we ask the right questions. It seems to be that we can match people to jobs agnostic of everything else. So pre covid, we got the gift from Microsoft, we got the gift from Syracuse University. We started using machine learning to build something called jobs ability. COVID hits. We’re not traveling speaking anymore, thank God. It’s a little bit easier on my body.

(31:19):

And we start building jobs ability. And as we do that, we realize that yes, right now today it takes eight minutes, eight minutes for you to upload a PDF of your LinkedIn profile, answer six questions. We’re matching you to jobs in eight minutes. The most important thing we’ve done is we proved our job recommendation engine works and is agnostic to other things that are external. And that was Microsoft’s whole mission to us as can you build it without the bias? And we can, in the process of doing that, a couple of things happened. One is we became a technology company without meaning to be. I think that’s the thing I didn’t anticipate. Be careful what you wish for. The second thing that happened is in the process of thinking about data sets, in the process of thinking about using it for good, our contact at Microsoft was a man named Kartik. Kartik never turned his camera on teams, never saw Kartik, but he was the person that I would ask questions to. And he was very stern, very abrupt, very point of fact. And so I would ask a question, he’d be like, yep, nope. Do this. Don’t do this. Yes, you can do that. Great. And I didn’t know what I was getting. I’m not like that. I like to talk a little bit. Obviously I’m sitting here, so there you go.

(32:41):

Kartik calls me in the middle of COVID and says, Hey, John, I just want you to know something. And I was floored that he called me directly. I didn’t know, and my bad for not doing the LinkedIn research on him and not doing the things that I should have done. Kartik called me one day and he said, Hey, John, I just want you to know something. He goes, what you’re building I needed, I’m a hundred percent blind. I was the first student coming out of India to go to Stanford a hundred percent blind. I needed what you’re building. Keep going. Don’t stop.

(33:14):

Okay. He called six months later and he said, Hey, John, I’m going to start my own company, or I’m going to come work for you. Which would you prefer? You start Tuesday. So I gave Kartik some ownership in the company, and Kartik and I are two people with disabilities that own Our Ability. And we do that because it matters very deeply that we are people for people with disabilities. In the process of doing this though, something that was just said about job descriptions a little while ago, we have this data set of job descriptions, and I’ve always wondered, why aren’t people matching? What is it about the systems that exist in place that are not working? And so we took a look at the job description in total, and we started matching our candidates intake into, like I said, within eight minutes to the job description. But it made us think about the job description. One conversation with the London SOC exchange, they asked us, Hey, have you ever looked at removing ableist language? Could you help us remove ableist language or look at ableist language in our job description? Well, yeah, that’s extremely important because I know firsthand there, I used hand and I have no hands. So there’s a little trigger. I know that yes, if there are certain words, phrases, terminology in a job description, it will stop me from applying flat out. Stop me.

Lex Huth (34:43):

Me too.

John Robinson (34:44):

Yep. Can we look at ableist language? And so we took the challenge from London Stock Exchange, and right now we’ve built an ableist language detector that is able to detect out ableist language on our job board, something that we’re going to be building with a couple of corporate partners down the road. So if you’re interested, interested, let us know. But that has led to building a plain language translation model. So for the NEURODIVERSE community or the IDD community, we’ve had people ask us, Hey, can you translate this document into a third or fourth grade language? Yeah. Pretty instantly we can. We’re doing this because we want to create a more inclusive workforce. And to do that, we have to make the barriers less to the application, to the job description, into the hiring process. And if I can show agnostic to disability that candidate X meets job description y, then at least I’ve overcome some of that hurdle before we’re there in that process. And so that’s what we do, and this is what I get excited about on a daily basis and talking about using AI for good because we can use it for good in doing a better job in leveling the employment playing field.

Lex Huth (36:07):

Right. So we did have a question next about ableism and job descriptions if you wanted to go to that.

John Robinson (36:13):

Sure.

Lex Huth (36:13):

Okay. So does anybody have any examples they’ve seen or any questions about ableism and job descriptions? We did hear a little bit about it yesterday in one of the panels. Come in with the mic. Oh, great.

John Robinson (36:30):

PEAT shirt.

Jerry (36:34):

So could most of the bars last night. Hi, Jerry, for the state of Oregon Employment Department, and one of the things I worked with employers is the must be able to lift X amount of pounds, must be able to twist and turn and squat and all that. It was mentioned earlier in the week that there’s accommodations. We could use lifts, we could use two man assistance, two personal assistance. How do you get around those where it may be a significant portion of the job though, without being ableist? I guess

John Robinson (37:08):

That’s a great, great question. So there’s two answers to this. When we take a look at ableism inside job descriptions, we’re actually flagging that. And we’re flagging it for multiple reasons. We’re flagging it because is it needed. So we actually, when we look at a job description and we see something like that, we’ll come back to you or to the employer and say, is this needed in your skillset of what the job entails? If it’s yes, no, then that’s the direction we go down. If you are a warehouse worker for UPS, you do have to lift 40 pounds. That is what’s going to happen. Yes, there may be some reasonable accommodation that can happen, but the truth is, I’m not going to apply for a job that is requiring me to lift 40 pounds. I’m not going to put myself in that situation, but I may apply for a job that says that if it’s 95%, that isn’t that job description. So when we look at ableism and we look at the terminology in it, we’re flagging that and then returning it back to you with suggested language changes based on if we see that job description and the rest of the job description really is about lifting weight or is it about something else, such as it’s an old job description that was always used. It was always used. The other one is the driver’s license.

Lex Huth (38:25):

That’s the one that gets me every

John Robinson (38:26):

Time. Yeah, every time.

Lex Huth (38:27):

Yeah.

John Robinson (38:28):

Must have driver’s

Lex Huth (38:29):

License, right? No, I should not have a driver’s license. Oh,

John Robinson (38:33):

Go for it. Go for it.

Lex Huth (38:35):

So it’s just one things too. It’s like, why would you require that of me as an editor? I have no idea. But you just don’t meet those requirements and they’re completely unnecessary.

John Robinson (38:47):

And that when we flagged that, we’re asking the question back, are you really saying you got to get to work on time? That’s really what it is. It’s not Can you drive or do you drive? It’s can you show up at 8:00 AM And that’s what we’re now, is it a CDL license for UPS, then yes, I hope you have a job. I hope you have a driver’s license. But those are some of the more tried and true things that we flag. But then there’s also see, hear, communicate, talk, those types of words that are in a job description that we’re flagging and saying, what are you really saying here? Good communication skills. Is that what you mean? Because communication skills for me and somebody else may be completely different. But if I’m a best editor in the world, but I may not be able to see, and I use voice dictation software, you know what? I’m going to hire you all day long because you problem solve and you work on it. So this is the modeling that we’re doing in looking at job descriptions to be able to break down the barriers, to be able to match people to jobs.

Lex Huth (39:48):

I also just want to mention too, a lot of people will say things like, you must be perfect or exceptional. That can be hard because people with disabilities often are already doing that so that nobody looks at us and judges us. But then at least I personally take the failures that I can’t control even harder. So one example of that is using speech to text and text to speech. Lightning like a lightning bolt and lightening, making something brighter, homophones, I can’t tell. And so I’m not perfect in that way, and we have to give each other a little bit more grace. If you see a typo, it might have been speech dictated, it might be something like that. So I see a lot of people being very harsh about it in a way that I used to be before I started to lose my sight in my twenties. We just need to remember that not everybody communicates or works in the same way, but that doesn’t make it any less valid. And they’re probably better at certain things that you don’t even know about because you’ve excluded them for something that doesn’t matter.

John Robinson (40:47):

Absolutely.

Lex Huth (40:49):

Oh, right here.

Hannah (40:52):

Sorry. Almost my down. Hi. I’ve just so enjoyed y’all’s presentation. My name’s Hannah. I work for UT Arlington, and one of the things that I have a hard time with my managers is they have this, I’m so glad that you brought up exceptional communication skills. Always like, what does that mean? I always look at them, what are you talking about? Tell me what you mean. And then the other thing that always gets me is able to work independently. I’m like, what are you really saying right now? What does that actually mean? Because you are going to exclude individuals who feel like they do need that accommodation, and they’re going to think, well, I can’t be 100% independent without it. But I guess my question is, do I really, what are the questions that you ask your clients to get them to actually tell you what they’re trying to test for?

John Robinson (41:38):

So when we do the ableism model, we’ll look through all the job descriptions at one time, and we’ll provide a spreadsheet. And in the modeling of the spreadsheet, we’ll flag ableist terms. We’ll make, we won’t make changes, but we’ll make suggestions. And then we’ll also score good bad type thing. And then we will talk with clients one-on-one verbally. We will not do it on paper necessarily, where we’ll go over each one and ask if you have questions and then make suggestions on if they ask me suggestions on terminology that we could change. Again, must have communication skills. What was the other one? You just

Hannah (42:23):

Able to work independently,

John Robinson (42:24):

Able to work, oh my gosh, that one. Yeah. Able to work independently. We’ll talk that through and say, what is it you’re really asking them to do? Are you asking them to be task oriented? Are you asking them to be time management? Are you asking them to be on time? Are you asking them to fill a specific job duty? If those are yeses, let’s put that in the job description rather than having an innocuous phrase like that. That could be confusing. And we don’t want to do that. As we’ve gone through this exercise, we’ve noticed a couple of things. One, because people are using ChatGPT for job descriptions. There’s some less, some more. ChatGPT throws more in there sometimes, which is very intriguing. But we’ve also noticed that the business community is more in tune to removing ableist language than they ever were before. That’s a plus. That’s a huge plus. And I take that as a small victory. But to answer your question, we’ll work with them directly. If they ask us on suggested language changes, I will never make a change to somebody’s job description, because you’ve got lawyers, you’ve got HR managers, you’ve got people you need to report to. You need to make those changes with them after I’ve given you the suggestion, but we will talk it through and I’ll give you my opinion and provable opinion on why we might suggest some changes.

Hannah (43:57):

And do you think there is a scalable format to do this effectively outside of the one-on-one, because that’s really where we are struggling as an institution. We have all of these who are just, there’s just a lot of ’em, and we want to be able to have that.

John Robinson (44:18):

We have Seth and I have an idea. Yes, there’s an absolute scalable model here. Awesome. So yes, if you’d like to talk after,

Hannah (44:25):

Please, yes, thank you.

John Robinson (44:27):

It’s important to listen. This is something we can fix. We can use AI for good to do this. We have the data set of job descriptions. We have 20 years of disability behind us collectively in work. I mean, obviously 55 years of being disabled, but we can do that and we should do it for good.

Lex Huth (44:47):

And I’d also like to make a note on just language in general, because when I started losing my eyesight and when I started talking about my OCD, everyone in my life kind of took a step back and they were so afraid to offend me or to say something wrong, and they stopped being themselves in certain ways because they were always checking themselves. It is about intention to, I mean, if somebody says, Hey, that’s not great, or That’s not the right term, or something like that, take it and don’t hold it inside. Just learn from it. Because I say the wrong thing. Sorry, I didn’t mean to tap my microphone there. I say the wrong thing about my own community. There are people who think I shouldn’t say visually impaired. It took me so long to come up with a term that worked for the fact that I might be blind later today or never fully. It’s a fluid thing for me. So I had to kind of come up with a middle thing. So just remember that nobody’s perfect. Meryl Evans always says, progress over perfection. She is a wonderful deaf advocate that you should all follow on LinkedIn if you don’t. And progress over perfection really fits here.

(45:54):

All right. So I guess we’re at the question and answer portion, but thank you all for asking questions throughout. It’s been very helpful. Thank you.

Eric (46:10):

Thank you so much for your presentation and sharing your stories. I’m wondering if you would be comfortable speaking a little more about your individual use of AI and technology.

John Robinson (46:23):

Sure. Which one? Both of us.

Eric (46:26):

Yeah, thank you.

Lex Huth (46:27):

Go for it.

John Robinson (46:30):

Well, I use FanDuel and DraftKings, so I’m using, I’m sure I’m using AI. No, very similar to what Lex said. I mean, I started using voice dictation software 20 years ago, and it was dragging nuance and speaking into my computer before it’s been standard in Apple and Microsoft. I certainly use that. For me, the biggest AI technology advancement is my cell phone. Wherever I left it, it’s somewhere, think about every app right now is using it to teach me to navigate the world. That’s the personal usage that I use, certainly. And I think about it more from the work experience world. I mean, if I were in the job market, again, I would be looking to use it, right? I’d want to use it to find a job to figure out the navigating job. In my instance as a business owner, can we use it to match to diverse suppliers in corporate America? Can we use it for other purposes? So I think about it from a corporate standpoint, but I know I’m using it all day long every day with my phone, with the apps that it’s using to help me navigate this world. It has made it easier. It does not make it easier for me to walk down these streets.

(47:57):

No, but it almost makes everything else easier. And that’s a plus,

Lex Huth (48:03):

Right? I think it was, if I got the voice right, Eric, who asked that question, so thank you. Was it, was I right? Yes. I’m really working on that orientation and mobility focuses a lot on hearing versus trying to see, and I’m going with that. But for me, using AI, I guess I should back up a little in my story too. So when I started, when my retina started detaching, they tacked ’em back up the best they could. And then they said, go out, figure it out. And so I was in grad school. I finished and started working, exhausting myself. Never learned a screen reader, never used assistive technology. No one ever told me there was no toolkit. And so at one point, my dad sent me a box of different technology, different manual things. It could be a magnifier, it could be a different monitor I could hook up to.

(48:55):

And he came back. He had been living in the Middle East at the time. He came back to Pittsburgh and said, you can do anything with this. We can take this anywhere you want to go. Let’s figure out how you work best. And so through that, I figured it out. It was a lot of manual. It was exhausting. And so having tools that make it easier, like AI when used inclusively and built properly, give me the ability to have that career back that I thought was going to be taken away, and that nobody really helped guide me to find again. And that’s true for so many people in so many different situations, not just with disability. Our lives shift quickly and nobody gives you a toolkit on how to fix that or how to change it or how to update how you think about it. So we need to build inclusion in so that we’re not all struggling when something sets us back or changes.

Hannah (49:51):

So my question is, how do you make the case for the preemptive strike of working on job descriptions and all public facing content when there are individuals in leadership positions who basically say, we have accommodations, we have that process, and we don’t have the time to go and do this. How do you do that? Such good question.

John Robinson (50:17):

Have you looked at Microsoft Stock price in the last 12 years?

Angie (50:22):

No.

John Robinson (50:22):

Astronomically through the roof because they’ve included with disabilities in the product set, in the purchasing and employment process. So they’re including people with disabilities at the beginning. So when I get a question like that from corporate America, I lead with that. How do you feel about your bottom line, right? And pretty every business, including me, I care a great deal about my bottom line. If you exclude people with disabilities at the language barrier, at the job description, at the product market, marketing, if you exclude because of language you’re using, you’re excluding one out of five people potentially. The labor numbers just came out this morning for the past month, and our labor market is getting better. Imagine how much better it would be if we added more people with disabilities to that component. The numbers that Lex showed at the beginning about buying power and the economics around people with disabilities, the buying power for people with disabilities in the United States alone is $1 trillion, and yet we’re 65% unemployed. If we went to 55% unemployment, real unemployment numbers, that 1 trillion becomes 2 trillion. So the answer I would give back to senior leaders afraid of even looking at language. Why would you exclude and create a barrier at the very beginning that is hurting, potentially hurting not only your employment roles, but your purchasing? Let’s be inclusive at the beginning so that we can expand at the beginning. And again, I would say look no further than Microsoft because every company wants that stock in the last 12 years.

Francesca (52:15):

Thank you, John.

John Robinson (52:16):

Yep. I

Francesca (52:18):

Have a question. What did you say? The buying power of people with disabilities.

John Robinson (52:23):

People with disabilities in the United States alone and their direct partner. So not kids. Not parents. $1 trillion.

Francesca (52:31):

Okay. So I just looked up African Americans we’re 1.6 trillion. So imagine if we stopped talking about all this dissecting and came together and collaborated. We would have so much freaking power. We’re 1.6 trillion.

John Robinson (52:50):

We just got

Francesca (52:51):

In the economy.

John Robinson (52:52):

We just got a request. Could we use our system for other purposes? Bringing it together? And I have to tell you, I’ve got goosebumps of saying that out loud. That’s exactly what we want to do,

Lex Huth (53:04):

Right?

John Robinson (53:06):

Yep.

Lex Huth (53:07):

We don’t want this to be siloed progress. We want it shared everywhere. This is not just about one group. It’s about all of us.

John Robinson (53:15):

Lex, thank you.

Lex Huth (53:16):

Thank you. And thank you all. Thank you everybody. I’m just going to put our contact information. I’ll just read it out real quickly. So John is O-U-R-A-B-I-L-I-T-Y, space INC, Our Ability, Inc. And I’m PEAT, P-E-A-T. That’s the Partnership on Employment and Accessible Technology. So thank you. Thank you.

DirectEmployers (53:45):

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