In today’s special live episode from NYC, we’re hosting a town hall meeting to discuss generative AI and its impact on talent acquisition. Our first remarkable guest is Jenny Cotie Kangas, a friend of the show and a formidable force. Joining us is also Dave Walters, the innovative Chief Technology Officer at Hired.
Rob Stevenson 0:05
Welcome to talk down to me, a podcast featuring the most elite talent leaders on the front lines of modern recruitment.
Speaker 2 0:12
We actually want to understand the themes of someone’s life, we want to understand how they make decisions where they’re willing to take risks and what it looks like when they
Rob Stevenson 0:21
fail, no holds barred completely off the cuff interviews with directors of recruitment VPs of global talent, CHR rows, and everyone in between.
Speaker 3 0:31
Once I went through the classes and the trainings and got the certifications through diversity and inclusion, I still felt like something was missing.
Speaker 4 0:39
Talent Acquisition. It’s a fantastic career, you are trusted by the organization, you get to work with the C suite and the security at the front desk and everybody in between and everybody knows you.
Rob Stevenson 0:52
I’m your host, Rob Stevenson, and you’re about to hear the best in the biz. talk down to me. Hey, everyone, I’m so glad you’re all here. Welcome to Talk talent to me live. I am Rob Stevenson, your occasionally humble host, we’re gonna have some fun tonight. Before we get going though, I want to do a quick poll. If you have ever been to another HR tech vendors, Happy Hour Meet Up panel event, say boo really loud. That’s what I like to hear. Hopefully, this will be nothing like that, we’re gonna have a little bit more fun. Tonight. Before we get going, I’ll tell you a little bit about what talk talent to me is all about. Normally, on the podcast, I sit down with directors of recruitment and VPs of HR CHR OHS these sorts of folks we sit down, we shoot the breeze, we talk about whatever they’re working on, it’s a ton of fun, it’s really casual. And that’s a blast, you should check that out. But today, it’s gonna be nothing like that. We’re trying to give you all something that you can kind of take home and work on back in your daily lives. And so it’s not going to be the Jimmy Kimmel type of talk show that we normally do for the live talk towns meet, we’re actually going to do some live demonstration and some workshops for you stuff that we hope you can put into your processes. As soon as you even walk into work or log on to work tomorrow, even so we’ll get into that. I’m sure you all are hearing about this AI rigamarole. And it’s everywhere. And there’s loads of hype, there’s loads of fear mongering, you’re hearing stuff like it’s going to take your job, it’s going to disrupt these industries, it’s going to make people irrelevant. AI is going to 3d print a black hole in your backyard and suck all the space and time into escaping infinite law and destroy life as we know it. Is that true? I don’t know. But crucially, it’s not going to help you get any better at your job. And so that’s why we designed this event tonight to just do just that. Our first guest is going to be the person who does that for you. I’ll bring her up in just a moment. But first, I’m just going to a couple of details about how we’re going to do it. We’ll bring her up, we’ll get to know her. And then I want to make sure that you all get a chance to ask questions of our panelists have a little more of a back and forth, we can turn this into a more of a town hall thing. If you want to storm the stage WWE style at some point. That’s welcome. As long as you have a really good question to ask at the end of it. That’s my only requirement. So we will get into it. Let’s meet our first guest. So I have to do a little bit of front coaching for who our guest is because she has an amazing story. She sustained a traumatic brain injury a couple of years ago and suffered complete amnesia forgot everything about herself forgotten her kids faces and her kids names, forgot how to speak and how to walk certainly forgot everything about her recruiting career up to that point. And so she had to completely relearn everything. And what has resulted in is her being as absolute shit kicking badass at work, who has no expectations about how things should work. He was constantly asking, why do we do it this way, not accepting. That’s just how we do it. And it’s been an amazing disruptive force. I have to tell you all that we simply cannot get into it tonight because we have things to do we have pawns to cast and whatnot. But Jenny and I did do a podcast episode where we talked about all this a few episodes back. So get out those phones. Subscribe if you want the full Jenny biography. We can’t get into it now. But you should check that out. Anyway, without further ado, is director of employer branding over at pandologic Jenny Cotie Kangas, Jenny.
Rob Stevenson 4:09
Hey, Jenny, have a seat. We’re sharing microphones because we spend too much money on champagne. So how are you?
I’m good. I’m good. Thanks for the intro. That was wonderful. Also, I’m a director at another HR tech company, just to be clear. So that was a fun little preview there.
Rob Stevenson 4:29
Was it one of the ones I told them to do when we were talking about live events?
I think you did it, which makes me a little sad.
Rob Stevenson 4:35
We’re off to a great start nowhere to go. But up, Jenny, I’m so glad you’re here. Would you mind for everyone here sharing a little bit about your background? We’re gonna get to the goods and the screenshare and the workshop and all that. But first, let’s learn a little bit more about you. Would you mind sharing about? Well, fast forward the good stuff that just that just previewed a little bit about your background and how you kind of wound up in your current role?
Yeah, so I wound up in my current role, because of that man who’s standing right there. Carry You Faker was absolutely incredible. So I had joined another vendor in the space. And they had gone through an unexpected reduction force. And Terry reached out and he said, come work for me. And I said, Well, I’m sorry. Like, what he’s like, No, come work for me. And so then when I challenged back, I was like, you know, I’m really wanting to stay in talent acquisition, because I’ve been doing a BPO marketing role. And Terry said, Okay, well, we’ll create a role for you. And he did. And so what I do at pandologic is I’m the head of talent acquisition, employer branding and awareness. And so one of the things with my story, which Rob kind of teased a little bit there, I do a lot of storytelling to the industry to help them kind of understand how to understand some of these concepts specifically, like AI, how do you leverage AI and automation to enable humans for the moments that matter? So we do a pandologic? And yeah, that’s how I ended up where I am today. And I’m really excited that our paths cross and then we’d have conversation today.
Rob Stevenson 5:54
Me to closure here. What are we going to talk about today? I guess we should get into how before we do the clicky, clicky? How are you conceptualizing AI in your role? And how are you making sure you and your team leverage at best?
Yeah, so like I mentioned, I’m a leader with a disability. So because of my head injury, I can only be on a screen for a certain like a lot of number of hours in a day. And if I get to the max of that I will pass out or have a seizure. So it’s kind of a hard guardrail. So what I have done is I leverage AI to enable me the human to be able to go further and faster at work. And so really, it’s a better question. It’s probably what I’m not using it for. But it helps me to project manage myself to schedule my day. Recently, it taught me how to boil lobster tails at home, which is actually really cool. I’m sure we’ll talk about that in a second. So I’m using generative AI a lot. And again, it doesn’t leave a lot of the mental brainpower that’s required to do a lot of different things. Like we have to switch so many tracks, especially in HR and talent acquisition, you’re having to wear all these different hats. And so really, I think of generative AI, or what I’m going to show you as essentially like my consultant that has all the hats. And so as long as I’m calling out the specific hat that I wanted to wear, a lot of the times I can get it to go like 75, sometimes 80% of the way, and then I can just sharpen it from there.
Rob Stevenson 7:17
This is an important call out that AI is not coming for your job, it’s more coming for the parts of your job you hate or in your case you physically cannot do right. And that leaves you to work on the things that only you can do that you uniquely bring to the table, which I think is awesome for all of us. Will you tell us more about the non work things though? So you’ve boiled lobster, you didn’t Google how to boil lobster, you your first reaction was to open up chat GPT.
Speaker 7 7:40
To be clear. My first reaction was to google how to boil lobster tails, which came up with hundreds of 1000s of different results and all these different recipes that were there. And I’m sorting through them. There’s all these ads. I don’t know if you guys have ever, like looked at a recipe online, but you should have to scroll all the way down to the very, very, very, very bottom. You might find a recipe. But then it’s like, the actual step by step is like embedded in there somewhere. And so I opened up chat GPT. And again, like I mentioned before, when you’re using chat GPT, you think of it as somebody who has all of the hats that exist there in the universe. In order to get it to respond the way that you want it to, you have to call it the hat that you want it to wear. So I said, acting as a an at home chef, leveraging ingredients commonly found in a kitchen. Can you tell me how to prepare lobster tails. And it did step by step by step. And I didn’t have to sort through the ads. I didn’t have to look at it. And it was right. We did the lobster tails that way. But that’s one of the ways that I’ve been leveraging it. I am moving recently got engaged. Exciting. Thank you. And so I am moving. And I’m also traveling a lot for work, because it turns out AI is kind of a big deal. And a lot of people want to learn about it, which is men, I do a lot of speaking engagements, which is great. But I have this new coming up. And so I actually had Chat GPT write a project plan for me to manage. And so I said I’m moving on this day. This is like what it obviously knew what the day was today. And I said, acting as a real estate consultant. Give me like a project plan to have a move. And also I have three kids so incorporate that in there too. And it did another fun way that I’ve used it. I had it write a card to my fiance. So acting as a Hallmark card writer. I want you to write a card to my fiance from my three children. And they called out the things that they wanted it to incorporate, which were wiffle ball. Because my fiance plays on wiffle ball team Yes, that’s a thing of Minnesota, as wiffle ball claw machines and puns. And then I added and I want you to do it in rhyme. And it wrote an entire card for me which like, I’m sorry, the hallmark for this No, but like, I think I’m ever gonna write like buy another card again, because it’s so great. Yeah,
Rob Stevenson 9:59
it’s certainly good. mean no worse than Hallmark at that point? Did your fiance call it out? Were they like, Hold on the Ryan’s name is way too tight? There’s no way you did this yourself.
Speaker 7 10:06
You know, it was me. So that was probably chat GPT right. And so part of the smooth, my alarm company wanted to charge me $1,200 To break my alarm contract, which I thought was ridiculous. And so I had Chat GPT write me a better business bureau complaint to this company, which it did in like 15 seconds. And it was incredibly well written. And I just copied and pasted it into the BBB complaint. And they dropped it on Amazon. And I got to keep doing the
Rob Stevenson 10:37
hardware to 1200 bucks. Not bad. Yeah. So yeah, don’t be afraid of this, folks. You can use it in all sorts of mundane ways in your life. AI is not just facial recognition technology on drones. It’s McDonald’s is also using it to make better ketchup, so don’t have it. We simply have to start the show.
No, no, no, no. So does anybody here have kids? Does anybody here have to help their kids with math homework?
Speaker 7 11:06
So for those of you who don’t have kids, math is very different from how apparently from how we grew up? I don’t remember it. But apparently it’s very different. Right. And so if you take the question to chat GPT, it will actually write instructions for you to teach it to your kid. Which is a really, really, really cool thing. So like I have, I mentioned three kids that are 1512, and five. And I have been teaching them how to use their powers for good with generative AI. So
Rob Stevenson 11:37
I wasn’t good at math the first time around. I’m definitely not going to be good at it when it’s time to teach tiny children how to do it. But any case and they’re a great example, Jenny, we should give these people what they came here for. Right. Should we get to some chat GPT? So Jenny has plugged into her computer in the back of the room, and she is going to just walk you through. Here’s some examples of prompts and things you can use in her job descriptions for hiring manager emails. I don’t know exactly what you’re going to talk about. I guess I just stop introducing it. But she is going to walk you through how it’s used in her daily work and sag so yeah, I’m gonna get off stage because this is very uncomfortable. If I’m just sitting here, but take take it away. And
Speaker 7 12:16
so what you’re looking at right now, if you’ve heard of chat GPT, that is a technology from open AI. And so there is like a paid and there is a free option for this. So Chat GPT 3.5 is I believe what the free version is 4.0 is available when you pay for it. So it’s $20 a month that I pay for chat GPT. And again, it helps me immensely when you pay for it. One of the cool things that you get, and I’ve cleaned these up a lot, but you can go back in to all of your different former things in the left here.
Rob Stevenson 12:47
HR tips for Vegas conference.
Speaker 7 12:50
Hold on. Where’s that? Oh, there it is. So what you see here is essentially chat GPT helping me sharpen my post. So it was five essential tips for an unforgettable experience. You can see this on my LinkedIn because I posted a sharpened version of this on LinkedIn. So what I had was a prompt. So we’ll start over here. So again for is the more advanced version. So I’m going to say here, I am presenting to a group of HR and talent acquisition leaders. And you can watch all of my different typos as I do this. So what I just did is I asked a sharp question, a sharp question, if remember if there’s all the different hats in there, so they give some context of like, what I’m looking for the use case, it’s going to be a lot more precise in its response. And so what chat GPT is doing right now is it’s explaining what generative AI is. So this is about how fast it goes. And really what it’s taking is, and we can go into this, but it’s giving you an answer. So generative AI is a type of artificial intelligence that focuses on creating new content. This is a different approach from other AI models, which typically categorize or predict based on existing data. Generative AI can create things like images, music, articles, or even human like Tech’s to understand how generative AI works, it’s helpful to know a little bit about the underlying machine learning models, and then it’s going into the specifics for that. But generative AI has a wide range of applications, especially in creative fields like art, music and writing. In the HR and talent acquisition field, for instance, a generative AI can be used to automatically draft job descriptions, create personalized recruitment emails, or respond to common candidate inquiries.
Rob Stevenson 14:38
Do you do any of those three things?
Speaker 7 14:40
So the way that I use chat GPT we’re going to give a little example. So can you provide 15 ways that this could leverage chat GPT in there day to day. So one of the things to remember, is any sort of generative AI, you have to ask the right sharp question to get the answer. But the other thing to remember is that you can ask the generative AI to be your consultant to help you figure out where you can use it. So here are some different options. It’s given me and this, again, was just like a basic one. For talent acquisition leaders, we have candidate engagement, automated responses, interview scheduling, job description, creation, candidate screening. So Rob, you had asked if I’m using this in my day to day, I use pandologic. And later today, so which is an AI enabled technology for doing some of these different things. So I don’t necessarily do it for those specific things. But what I do do is I have it builds my project plans for like how I handle the day, or if I need to respond to an email, I can take the email, that was the first email and I can say, write a response to this, which helps a lot. Another thing that you can do is if you work for an organization that has any sort of content that you’re putting out there, you can copy and paste the content into chat GPT and ask it to assess the tone of voice. And so it’ll give you the specific indications like what the tone of voice is, all sorts of different things.
Rob Stevenson 16:14
That tone of voice would be like, is this using exclusive language or discriminatory language? That kind of thing?
No, that’s what you need, like a text to or something for. So that’s another good thing to remember. Generative AI. So what you’re looking at right now is it was trained on everything on the internet before 2021, we have to remember that the things that were on the internet before 2021, a lot of them were biased, right. And so it’s going to say, based on history, this is what we think good looks like. And then it’s going to write that if you’re using it for content generation, it’s really important to have some sort of bias checker layer on there, like a textarea was another company that exists, that’s going to help make sure that you’re not biasing people at scale. Cool. Great. Awesome. Another really cool way. Has anybody here ever had to ask for more money in their job? Cool. A really cool use case for
Rob Stevenson 17:05
chaturdashi from people to their bosses right now.
I’m looking at the back of my boss’s head right now. Yes, so one of the things that you can do is you can ask chat GPT to help you write a better business case. So that’s one of the universal things that’s often I’d say, not as sharp like on the HR talent acquisition side. And it’s hard to know where to begin with that, you can bring in a situation, which I think it’s gonna take too long to go through it right now. But like, I could bring it a specific situation, let’s say, for example, that my job board advertising, I’m not getting the candidates that I need, right? So if I have to go to my leader to say, hey, I need more money, I can give it the use case. And then I can say, this is what I need, can you help me write a compelling argument for that, and it’s going to help give you those different options. You can also say, Give me three different examples, and it will actually give you things to choose from. So that’s one of the different ways do you want to take any questions like from the audience, like?
Rob Stevenson 18:04
Sure, we could uh, does anyone have questions for John, at this point, we could bring in someone from the crowd, if you’re a burning question you’d like to chip in with,
or a use case of like, how would you do this? I’m happy to like, just put that cool.
Rob Stevenson 18:17
I’ll come over. Hi, there, one brave soul. What’s your name?
Hi, I’m Amy. So my question is around what you said, with like, asking sharp questions. So what are specific? Like you should not
Speaker 7 18:32
perfect? So I’m gonna give you an example. What is AI? This is a non short question. Okay. So it’s gonna give a very, very broad response. It is not specific to my individual use case, it does not who know who I’m looking for. So again, like I mentioned before, think of generative AI as having all the hats that are out there in the universe. In order to get it to respond more precisely for your specific use case, you want to call out the specific hat you wanted to wear. This might look like, for example, so I recently stepped into doing some work with strategic alliances and partnerships. It might look at what are the different categories of partnerships, what are the different definitions of the categories of partnerships, and I had prompted it first with like, here’s what pandologic My company is. And that actually builds everything out with examples of that, does that help answer your question a little bit. So that’s what is AI, right. So as a talent acquisition leader, who is considering so this is a sharper question, where I’m saying, as a talent acquisition leader, who is considering leveraging chat GPT in my day to day, what are some things I should consider about choosing to begin? So now it’s giving me a bunch of different options? Has anybody seen prompt packs, like on the internet, where you can pay for a bunch of different things? So a prompt Here’s the question that you put in to that textbox at the bottom in order to give the precise answer that you’re looking for. And so one of the ways that you can do that is to buy a prompt, Pat. The other thing that you can do is have chat, GPT act as your prompt engineer and provide you with a specific prompts to be able to respond to something. So I will do that as an example, here in a second. But you’re seeing a much like broader answer that’s more specific to the use case than what it was at the beginning. Because again, I have to narrow its focus. Otherwise, it’s going to give me the blanket response. And there are really, really, really awesome insights that are in here. But in order to get that You’ve almost got to like play battleship, to get it to ask the right questions, so that when you throw it out there, it’s gonna hit any other questions. There’s one recommender next to you, Rob.
Rob Stevenson 20:50
Hi, there, what’s your name?
I work get hired. So this is cheating a little. But I want to know for my own use case, are there too many inputs? You can give? Chat GPT for a prompt. Is there a limit? Is
Speaker 7 21:03
there an ideal number? Are there too many inputs? I can give Chat GPT as a prompt,
Rob Stevenson 21:15
why make up an answer, right? We have the source of all knowledge and God itself right in front of us. But you
think it’s not God? Right? Like, this is AI. It’s AI that has like, it’s calling from a bunch of information. But here we go. So chat GPT three, which was the legacy version, the limit is 2048 tokens, which includes input and the output of token can be as short as one character, or as long as the word. So you can input a long prompt, but the longer the prompt, the shorter the generator response will be before hitting that token limits.
Rob Stevenson 21:48
So you think you can squeeze it into 2048? Their Alexa? You’re reading blog posts and Jegede. Right, well, okay. Anyway, that’s helpful. Jenny, this has been amazing, we do have to keep the show moving. It’s a tight operation we have here. So let’s keep this show on the road. I get it for Jenny, everyone. Danny wants to come back up here because I’m gonna want to bug you in a minute. Once our other guest is, is here. So come on in. I’m gonna slide you down the bench. But that’s okay. Just to reflect a little bit on what we just saw. I wanted to bring in a tried and true technologists to chat a bit about where he sees this application going. How easy is it affecting recruiters and an AI writ large? This man has had several technology roles at companies like the not XO group, and others. He’s a stand up guy and a volunteer firefighter as well. He happens to be hired as Chief Technology Officer. Give it up for Dave Walters. Everybody. Welcome, how are we doing?
Dave Walters 22:47
Good job. Sorry, let me just finish emailing my boss for my race.
Chat too busy to write your email for you?
Rob Stevenson 22:55
Oh, no, that’s that’s just for the content afterwards. This is the magic one that I have in my hand. Right now. I’m gonna hand it over to you in a minute. Dave, I just love for you to react a little bit. Was that surprising? Was that about what you’d expect to see someone in the talent roll using this for just softball question. Well, how did that make you feel? therapy session?
Dave Walters 23:15
Yeah, I mean, all the above, right, that I think this is such a new technology, right? We’re all constantly learning the new use cases that everybody is creatively coming up for. So I think that’s a great example of some of the core things that we can use chat GPT and generative AI for in our daily lives. Yeah, I mean, I think we’re gonna continue to learn, think back to for those of us that were around, I won’t look at anybody specific in the crowd here. When Google and all these other services were first coming up, right? A lot of the early days was just learning and figuring out what we can do with it right, how to ask the right questions, how to get the right responses, right now, you look at Chat GPT. And that just seems like ancient technology, right? But we’re gonna see the same kind of patterns and learning curves and evolution. So as we all become experts.
Speaker 7 24:04
So one of the cool things about chat GPT versus Google is I kind of showed you that there. You can prompt it, to teach you how to use it. And then you can ask it to give you the specific prompts to get it to act in that way. So we didn’t do that, because it takes a little bit of a long time to generate those things. But let’s say for example, one of these use cases up here was what you would want to be able to call it for, you could say for number five, or for this specific one, what is the specific prompt I would need to use to get you to respond that way. And so chat GPT can actually act as your prompt engineer, which otherwise would have been really a hard thing to kind of find in Google. So it’s a easier user experience, which is nice.
Rob Stevenson 24:47
Dave, if you found out your team was using this to do 80% of their role would you be like which we mad or would you be excited? i My question is like, is this naughty? Or should we all be trying to use as much as possible.
Dave Walters 25:00
Now another great question. So I assume that as every week goes on now, my team is using it more and more, right? I know, definitely they are. In certain cases, it’s very clear they are in certain cases, they’re telling me they are. But like I said before, we’re just finding more and more opportunities and ways we can leverage it to make our lives more efficient. So I embrace it. Right. I think a lot of people are embracing it. There’s some cautions, though, right? And, and we are trying to get ahead of that with my team within our company and setting policies and setting the right procedures in place for when it should and shouldn’t be used what you should and should not do with it, right? Don’t upload all your business secrets into chat GPT, right, that’s probably not a smart thing to do, I would avoid doing that. But you know, you can get some more generic information back from it. We’ve used it to start to build out new policies right from scratch, right, get the beginning of a framework there. And then we can evolve it from there, those things are generally safe to do, right, because it’s drawn from public information that is out there, I would recommend to everybody, if I don’t know if some of you are in the tech, field tech leadership, make sure you’re getting policies in place, you’re setting policies, you’re communicating those policies to your teams, so they know what to do and what not to do. Don’t let that policy be, don’t use it, right? Like that is a bad approach, right? You are going to fall behind, it’s inevitable, people are going to use it anyway, behind your backs. Even if you do say don’t use it, but just find the safe ways to do it. And the way to do it, the ways that are going to benefit your employees and your team members and your company.
Rob Stevenson 26:36
I don’t know if you’ve been to an HR tech conference in the last 10 years, but AI has been everywhere for a while now. And a lot of those cases, it’s kind of just been marketing and hype. All hat no cattle, as we say in the boring I state from which I help, but it feels different now, is it because of generative AI, it feels like there’s been kind of a Rubicon moment, do you find that? Do you find that we’ve moved past a little bit of the hype, and now we’re into the real world.
Dave Walters 27:02
So I think what chat GPT has done is bring this practice in a way that all of us can really understand what it’s doing, right? This has been around us for a while now. Right? It’s been around and a lot of the tools we’re using in our smartphones and a lot of services we’re using online. But it hasn’t always been very clear that we’ve been using AI right. A lot of it has just been behind the scenes, augmenting the things we do, right, even simple things like sending a text message, replying to a text message and getting those helpful suggestions of what you may want to respond with, right. And a lot of my messages to my family members are just clicking on the pre canned response. That stuff is all related. This is now going a step further. Right. It’s making it very clear, very intentional, very much in our face that this is what’s happening. So we’re connecting the dots a lot clearer. Right. And that is the significant transformation that’s happened in the last six months or nine months, I guess now,
Rob Stevenson 28:00
Speaker 7 28:02
sweated, like, continue, I think it’s democratizing AI, it’s creating a more level playing field. And if you use its powers for good, that can be a really great thing. For example, like if I needed to navigate, let’s say, applying for social benefits, or like applying for medical benefits through my state, right, that pass unless I knew somebody is typically a really, really hard one to be able to navigate, right? How about like a lot of the government programs that exist? A lot of the times they’re written like for people who understand the government understand, but not for a layperson, chat GPT generative AI can help you write more consumable content. So I didn’t show this here to you there. But like, at pandologic, we’re a tech company, right, like so we’re an AI company, that kind of is like a hedge fund for you, but your jobs. And so in the back end, like all the nitty gritty of that is very technical. But I have taken the here’s how we do it, and said write this in consumable, understandable terms for somebody who doesn’t understand tech, and it will rewrite it for me. And so like, that’s another way it’s like, so it’s making a more level playing field. But again, you’ve got to use kind of your powers forget.
Dave Walters 29:15
Let me just say to check GPT that’s what we’re all hearing about. Right? That is like the face that’s in front of all of this, but it’s not the only thing out there, right. I’m sure a lot of you are using Grammarly at work right to help with our writing. I’ve been using it for a long time now save me on a lot of embarrassing spelling mistakes, that also has some generative AI built in now, if they launched it fairly recently, it’s been a little bit quiet, but it’s open to everybody even without a paid subscription. And you can have it respond to a full thread. It’s looking at that thread looking at all the context of it. You can give it a couple of cues and it’ll spit out a whole email for your full email response. So it’s there it’s around it’s it’s more than chat GBT and it’s gonna keep growing Oh, yeah,
Rob Stevenson 30:01
yeah, good call out autocomplete on yours. Do Google search or writing an email? That’s a language learning model. Or if you bought your camera and the little squares around faces, that’s a wee little algorithm running right there on your, your baby phone. So it’s everywhere. Don’t be afraid. It’s just happening in your life. And you can use it for your own good, I think is what we’re trying to get at. Jenny. Yes, please. This is your show. Now I don’t even really need to be here.
Speaker 7 30:25
One of the other things that I’ve found it to be really helpful for changing legislation happens a lot, right. And a lot of the times when we have something that that happens, for example, like the Colorado equal pay app, that says that we have to put the pay, and then the benefits on a job advertisement, if I’m looking at that with my talent acquisition strategy, as just Colorado and missing the bigger picture, right. So what I can do is I can take the block of like information about what the Colorado Equal Pay Act is, I can take CCPA, a little block about that. I can take local law 144, which is coming out of New York. And I could ask Chat GPT to give me the meta implications of these laws. Right? So to scale back, instead of looking at the really, really tiny, like aspect of this specific piece, what are the meta implications of some of these things. And that is what we should be using to build their talent strategies, not the specific like the little use cases. So it also can help you kind of go really, really high back, but then kind of get laser focused in and then be able to build something that’s going to work with all those things. And
Rob Stevenson 31:30
can we speak about the New York statute that’s submitting people for an audit? If you use AI? What is the number?
Rob Stevenson 31:36
What was it called
local law? 144?
Rob Stevenson 31:38
Alone? 44. Okay, maybe you can you give like a quick high level for folks? And then under what circumstances should someone be concerned that this thing is coming to eat their lunch?
Okay, let me preface this by saying because my boss is sitting in front of me, I am not a lawyer, you should make sure to talk to a lawyer about this. But essentially, local law 144 is legislation that’s coming out of the city, New York, where we are right now, it was passed, and it was put into place, January 1 2023. So this year, and fourth minutes and start until July 5. What this law says is that if you are an employer that is recruiting, in the city of New York, you have to undergo a bias audit, to be able to understand if your technology that is in your experience is biasing people at scale. And that audit has to be third party. So at pandologic, what we did is this applies to employers doesn’t apply to vendors like and so at pandologic, what we saw is like we wanted to get out in front of this because we understood that, like employers weren’t always going to understand how this all worked. And so we went to a third party group, we opened up the door to all of our different algorithms that we have in the in our experience, we essentially said go fishing, find anything that you think is gonna apply, that could potentially bias people at scale, they found two algorithms that they thought would fit the bill. And then they went deep and wide on those two algorithms to make sure that that isn’t impacting people at scale. So the cool thing was like, what we found is like, our technology actually enables having a more diverse type of button, which is cool. But then the next step was we’re working on like these consumable action guides for understanding how as an employer, you need to start. So if you are an employer who like is recruiting in New York City, right? Again, this goes into place, July 5, which is soon. And there are massive financial implications if you don’t do this, right. So it’s really important to look at it. But the thing to remember too, is like if you do any remote hiring, and the candidate might actually be in New York, it applies in that situation as well. But figuring out like, what’s your strategy for going into? And how do you begin? The first step, it’s always critical for anybody is you’ve got to map your processes from high to higher. So like, what are the actual candidate journey processes that you go through? Because in order to understand its automated decision making tools being used, you need to know what your process is. So then once you have it mapped, you can actually figure out, okay, this is where automated decision making tools are being made. And then you can go one step further with talking to your vendors, for example, to make sure that the technology that you’re putting in place doesn’t bind people at scale. I got technical on that. I’m sorry.
Dave Walters 34:16
No, that’s perfect. And all that there’s obviously there’s financial implication to this. But there’s also social implications. So that’s right. So I think we all want to be responsible with our AI. And that’s what it’s all about. This is something that hired has been very focused on as well. We underwent a third party audit last year. We got early on the bandwagon and just like, like you did, we wanted to make sure even before we were even clear if we fell under that legislation, we wanted to make sure that we were doing the responsible thing, right that we did not have any bias within our algorithms within our platform, because ultimately, that was going to benefit both our candidates and our clients, right. And we’re uniquely positioned to where we have to do right by both of those constituents.
Rob Stevenson 34:58
Bearing in mind that either of your lawyers, if you submit to the audit and you fail, do you then sustain the penalties or is like, Okay, now here’s your path to compliance. And we’ll check in with you again next quarter or something.
Dave Walters 35:12
So I can ask chat GPT as a lawyer
to reset, but
Dave Walters 35:19
So my understanding the legislation was very clear on you needed to have had the bias performed before this point in time. There’s a lot of points about how you need to operate. There’s a lot of gray area, like a lot of new legislation, there is a lot of gray area, it’s going to have to evolve and be refined over time. I think there’s a little bit of intention there with the legislation where they maybe wanted to be much more clear and much more specific, but realize it was too early, too fast, right? They needed to be a little looser at the start start to get companies trained and used to operating this way. I expect there’s going to be a lot stricter guardrails in the future.
Yeah, and the thing to remember with a lot of this is right, we as HR leaders, and talent acquisition leaders, we’re familiar with this world, often that people who are writing the laws aren’t right. And so sometimes things are gonna come in place. And it’s like, you can look very, very specifically at the law. But again, if you’re able to pull back, what are the meta implications of this? So for example, there are two pretty known cases about bias that a lot of these laws have reference, one of those came from a company called Amazon, that was leveraging technology that essentially started training on what good looks like for internal promotions. And because it was modeled after humans, that pattern, right, reinforced by his behavior, which is a bad thing. And there was female, Latina women were actually really like, actually affected by this. There was another one that came from a company called Facebook. Are you guys familiar with that company? Cool. So they had a job posting, I guess, product at one point where you could specifically target the age range of the candidates that you wanted to apply. ageism is one of the biggest biases that exist in our world today. That is a no, no, you don’t want to do that. And so like, again, these laws at a very, very high level, are meant to ensure that we’re using technology for good and that if we’re leveraging any sort of machine learning, we’ve got guardrails in place, so you can understand what’s going on pulling back that curtain. This is a new terrain, right? Like this is something that we haven’t necessarily had to handle before. But there are technologies that are out there that can help you learn the right sharp questions to ask a vendor. And like I mean, hired is working on this too. If you’re working with a vendor, and they won’t open up their door, or they won’t give you answers in this. This is what we call a red flag. You don’t run towards red flags, you run away from them. So it’s another case is just kind of teaching how to play the game, right?
Rob Stevenson 37:53
Tech compliance advice and dating advice from Jenny, could you can guess you’re running away from red flags? How about that? We are creeping up on optimal podcast length, but I want to make sure we get to meet a couple other folks out here. Are there any questions for our panelists? I’d love to take take let me have a hand over here. I’m gonna walk over here and see what you have to say. Hi, there. What’s your name? Sorry. During Goldman, a friend of the podcast who has appeared on the show, check out our episode, and she’s slightly it’s not getting near the backdoor. And good to see you again. Hi, nice to see you. What’s your what’s your questions?
Speaker 9 38:23
So basically, given chat GPT is pulling from so many sources of data and whatnot, and we’re working in a very regulated and like, you know, no auditing and everything is being audited and measured and, and whatnot, who was verifying the data that’s being presented to us? And that, how do we know like, when we’re putting in our questions or information that is coming back to us, is accurate? And, you know,
Rob Stevenson 38:50
great question. I’m getting my stuff sitting here. This is fantastic. I’m creeping up on 10,000. I can feel it, I want to take the sake of this one.
Dave Walters 38:57
So at the end of the day, you are responsible as the user for validating that stuff. There are some you can probably see it above me, I think it’s right, somewhere over there. It states that the information it sends back may be flawed. I don’t remember the exact terminology. You can read it. You’re right. The data that’s being fed into it has some biases to it has some some flawed information in it, it is only able to operate off of what it has been given. Right. So it’s not smart enough to figure out what is factual and what is not factual. It’s today. It’s a big problem, right? It is one of the biggest challenges with generative AI, it’s going to have to be solved. It’s not easy to solve, right? Because someone’s going to have to make decisions of what information it should and should not get. I personally don’t want to be that person. That’s going to be a tough job to have. But in the meantime, you have to audit whatever you’re using there and make sure that it is fit for consumption.
And I would say it gets me 60 to 70% of the time, right? So it’s like it’s really the mental calories that I otherwise could have put together that email or I could have, like, I know how to do those things. But if I can get 80% of the way, and then I go and sharpen it, that looks good right there, that helps me say, in my situation with my disability, it helps me save a lot of mental calories that I can redeploy to engage with my team, or to working on different projects or attending events and things like that. So it really gives you a really good starting point. Yeah.
Unknown Speaker 40:31
And I’ll just add, obviously, we live in a very litigious society. This is a very high profile tool. lawsuits are going to come right someone’s going to try to sue Chat GPT someone is going to try to sue somebody that leveraged it to produce content that was flawed. You don’t want to be that that person. I’m not a lawyer, but you don’t want to be that person.
Rob Stevenson 40:53
Any other questions out there in the in the field? Yeah, here’s one. Hi, there. What’s your name?
Rob Stevenson 40:58
Hey, Kelsey, where do you work?
There’s the streaming.
Rob Stevenson 41:01
Disney. I love Disney. Plus, thanks for being awesome. We should We should talk later on when your podcast so sweet. Oh, thank you so much. I can live for like three weeks on a compliment like that alone, but I do appreciate it. What’s your question, Kelsey?
So what type of pushback Have you gotten from Legal to Use the chat chat? GPT?
Rob Stevenson 41:19
Jenny’s laughing. So I hope he answers a lot of my moments. I
Speaker 7 41:23
can go into this. But I other half maybe a lawyer for a very large company? And does their SEC security law? I can’t say the brand. But we have a lot of conversations about this. Because there are not guardrails that exist today. Right. For a I think the largest area that you will see in the law world is going to be aI regulation. So some of the things that are being said, from a legal standpoint, it violates your information policies, like what’s the specific word like information use policies? Yeah, so essentially, like if you give something that’s proprietary to your company, to chat GPT, it’s now open source, technically, right. So we don’t know if somebody can actually pull that specific thing that you put in there. But like, again, that would if monster is a big one. So but this is also being this type of technology. So large language model, which is what this is based on, is being embedded in things like Microsoft and things like Google, right? So I think you’re gonna see domain specific large language models, like we have a domain specific large language model that pandologic that we use, but a lot of different things that are coming from legal. And again, anytime something’s new, it’s a lot easier to attack it than to try to understand like, how can we put effective guardrails in place so that we can use this for good?
Rob Stevenson 42:41
Thanks. So Well, I think we are going to have to bring this podcast lighting into home here. So maybe I’ll have time for one more question. I want to ask both of you. What is the best career advice you ever received?
Dave Walters 42:55
All right, I think I heard it today. Ask your boss for a raise. Now, be flexible, right, be open to new opportunities. We were talking about this a little earlier. realities, I got where I am today by just exploring paths that I had no idea I wanted to go down right? opportunities were there. I jumped in with both feet, and worked hard. And if you’re open to new opportunities, if you work hard, amazing things are gonna come to you.
The hard question, I guess, best career advice, questions or your pickaxe, they help you figure out what the problem is. And so figure out make sure you’ve got really good sharp questions that are in your tool belt, so that when you come up to a problem, you can ask the sharp questions to figure out what’s really going on. Because often we have it designed by us or what we think’s going on. And sometimes we can miss the specific like, right so like questions are the best tool you’ve got in your tool belt, and make sure that you’ve got really sharpens.
Rob Stevenson 43:55
Well, folks, we’re going to be wrapping up here now but we’re going to be hanging out for a few more minutes, there’s plenty more to eat and drink. If I see a single champagne flute full in the next hour, the podcast is cancelled. Also, we are still doing headshots. So meet our photographer, they’ll take a photo of you and then we’re going to be providing these AI generated photos of you that you can use LinkedIn and elsewhere want you to have that for your little digital goodie bag before you leave. So make sure you take advantage of that and we’re gonna be hanging out so if you have any other questions for our panelists come say hi, have a drink and hang out with us. Until then. Jenny Cotie Kangas has been Jenny Cotie Kangas, Dave Alderson and Dave Walters I’ve been Rob Stevenson. You’ve all been amazing wonderful recruiting darling monsters have a spectacular week and
Rob Stevenson 44:44
talk talent to me is brought to you by hired hired empowers connections by matching the world’s most innovative companies with ambitious tech and sales candidates with hired candidates and companies have visibility into salary offers competing opportunities and job details. Hired unique offering includes customized assessments and salary bias alerts to help remove unconscious bias when hiring. By combining technology and human touch, our goal is to provide transparency in the recruiting process and empower each of our partners to employ their potential and keep their talent pipeline full. To learn more about how we can help you find your next great hire, head to hire.com/tt2m.