Is AI Taking the Human Out of the HR Department? – WSJ's The Future of Everything – WSJ Podcasts – The Wall Street Journal
Will the human resources department be replaced by robots? Not quite, but the use of generative artificial intelligence in HR is on the rise. WSJ reporter Chip Cutter tells us how companies are incorporating AI tools internally and what might change in the future. Plus, we hear from Reshma Saujani, the founder of Girls Who Code and Moms First, who recently introduced paidleave.ai, a free AI-powered chatbot designed to help workers navigate paid family leave benefits. Saujani tells WSJ’s Charlotte Gartenberg about what she sees as the potential risks and benefits of AI in the workplace.
What do you think about the show? Let us know on Apple Podcasts or Spotify, or email us: FOEPodcast@wsj.com
Further reading:
New York City Passed an AI Hiring Law. So Far, Few Companies Are Following It.
How AI Will Change the Workplace
HR Departments Turn to AI-Enabled Recruiting in Race for Talent
The Do’s and Don’ts of Using Generative AI in the Workplace
This transcript was prepared by a transcription service. This version may not be in its final form and may be updated.
Charlottte Gartenberg: Perhaps you or someone you know has exclaimed in dismay, "AI is going to take my job."
Speaker 2: Critics fear the rapid growth of AI could threaten jobs or be used for malicious purposes.
Speaker 3: Another concern is, could AI take jobs?
Speaker 4: Let's focus now on AI's impact on the way we work. One of the biggest worries people have about AI is, will it take our jobs?
Charlottte Gartenberg: It turns out some companies are already using, or considering using, generative AI internally. According to a Gartner survey from June 2023, most organizations appear open to adopting new HR tech. Only 15% of the HR leaders they surveyed said they had no plans to add generative AI to their HR processes. So whether you're an AI optimist or pessimist, generative AI is likely coming to your workplace, if it's not there already.
Chip Cutter: A lot of CEOs are going to their teams and saying, "Okay, you work in marketing, how are you going to use gen AI to sell better, to operate faster and cheaper?"
Charlottte Gartenberg: That's WSJ reporter, Chip Cutter. He covers workplace, management and leadership issues for our corporate bureau.
Chip Cutter: Some companies, for example, say that they think they can give up doing photo shoots for their products. Why couldn't you just have various sort of AI tools do that for you?
Charlottte Gartenberg: From the Wall Street Journal, this is The Future of Everything. I'm Charlotte Gartenberg. Today we're looking at our AI HR future, how close it really is, and what sorts of generative AI resources might become commonplace for human resources. Stay with us. AI and HR has been a thing for a while, but it's getting more widespread. We reported in 2018 that nearly all Fortune 500 companies were already using some form of automation in their hiring processes, whether that's robot avatars interviewing job candidates, or computers weeding out potential employees by scanning keywords in resumes. Last month, WSJ reporter, Chip Cutter, went to the World Economic Forum in Davos, Switzerland. After talking to CEOs and company leaders, he says that we're likely to see lots more AI in human resources very, very soon. Chip, what have been the pros and cons of this kind of AI use in HR?
Chip Cutter: Well, it's helped a lot of companies be faster, that they've been able to do more with less. For example, if you're screening millions of resumes at a large company, you probably need automation software to help you do that. Otherwise, it'd be really hard for one person to sift through all that. So that's been around for years, right? It's like ways to figure out the right candidates to interview. Ways to, for example, look at who might be at risk of quitting. Companies oftentimes have built complex software to see who's been in a role for a given amount of time, who maybe is in need of a promotion and hasn't got one. There's all these ways that companies can kind of get a sense for whether someone's a flight risk. So that's been one way that companies have used some of this technology in the past. But right now, I think we're on the cusp of a lot of change. And many companies, many CEOs, are telling their HR chiefs, and all of their department heads, you need to think about how gen AI is going to change how we work.
Charlottte Gartenberg: You were just in Davos. So I want to talk specifically about generative AI because it's being more integrated into workplaces. What have you been hearing?
Chip Cutter: You could not walk down the Promenade, the main street in Davos, without just seeing one display after the next about how AI is going to transform corporate America. Every interview with CEOs would somehow come back to generative AI and how they were trying to use it within their companies. Some organizations have a real plan, others are sort of earlier in their efforts. But a lot of CEOs are going to their teams and saying, "Okay, you work in marketing. How are you going to use Gen AI to sell better, to operate faster and cheaper?" Some companies, for example, say that they think they can give up doing photo shoots for their products. Why couldn't you just have various sort of AI tools do that for you? But then AI and HR was a big discussion as well. So for example, one HR chief at a large tech company told me that she has done an interesting experiment with gen AI. And there was a case in her company where a manager and that person's direct report were not getting along. And so she said, with both people's permission, can we upload the chat logs between them and see what happened? Why aren't they getting along? And so they ended up uploading pages and pages of logs, and they came up with some sort of interesting answers. And one was that the employee was asking way too many questions, and then the manager was getting frustrated by this, and the manager also felt that the manager wasn't being heard. And so just having that information, they were able to go back to these people and they said the relationship actually improved afterwards because both sort of knew, okay, this is what's bothering the other person. So it's almost like AI as corporate psychologist.
Charlottte Gartenberg: That's wild.
Chip Cutter: That, to me, was the AI example that I remembered the most. And that stood out to me as the most distinctive in a week full of conversations on this. But to me, it shows that HR chiefs are thinking, could gen AI really change how we operate? And if you think about this example, it'd be really boring for a person to go through pages and pages of logs and try to see, wait a minute, why do these people not like each other? What's happening there? But a machine can do that pretty easily.
Charlottte Gartenberg: But who might have an objection to having AI play that role in HR?
Chip Cutter: Well, you could see workers being frustrated if this is done in a case where they don't know about it or they haven't given their permission. And if we're using corporate chat tools, we've given up our permission. Our company is able to sift through that. But in this case, everybody knew this was going on, but you could see, down the road, where some might think, this feels a little bit like Big Brother.
Charlottte Gartenberg: What have you been seeing in terms of more high level attitudes about generative AI in HR?
Chip Cutter: Yeah. So many HR executives are optimistic about it. They do feel like they want to use it to some extent. And it's oftentimes cases where companies might roll out, for example, a chatbot that helps people. So an employee could more easily ask, how many vacation days do I have left, or what is our policy on X? The type of questions where someone might go to an HR person and ask them, some HR people think like, why can't we just build a tool that answers them more easily? That's one area where a lot of looking, uploading policy manuals, uploading benefits guides, all of this stuff where it just might be easier just to have gen AI sort of be an aid that helps employees navigate the company. Overall, HR, like a lot of functions in corporate America, is being asked to do more with less. There's companies… this is this, as we're continuing to see, sort of an era of layoffs and white collar job cuts in particular, so it's not that HR teams are going to get tons of additional people this year. So if you're in an HR executive, you're trying to figure out how do we do more with less? And many are thinking gen AI could be helpful here.
Charlottte Gartenberg: So what stands in the way of there being more generative AI in HR in the future?
Chip Cutter: Part of it is cost. I mean, companies might think they'll save money with this eventually. But not in the short term, it's going to cost money to set this up and to be able to do this. So that's one immediate barrier. And the other is just technical know-how. Companies are at all different stages on this. I mean, some executives are really immersed in it. Some companies, for example, have rolled out sort of mandatory gen AI training where for example, by March, everybody in the company has to complete this multi-hour course on gen AI. So you're up to speed on it, you're experimenting with it, you're thinking about how it could impact your job. And then there's a lot of other companies that are saying, "This is still hype, it's still unproven. We don't want to be an early mover on this. Let's hold back." So it really, I think, depends on the attitudes that executives have towards this technology too.
Charlottte Gartenberg: Yeah, I got that impression a bit from this Gartner survey from this past summer, that found that only 5% of HR leaders were reporting that they were already implementing generative AI. And only 9% of those surveyed were currently conducting generative AI pilots, which seems low to me.
Chip Cutter: Very low.
Charlottte Gartenberg: Is this going to take off soon, or… ?
Chip Cutter: You keep hearing that. Well, 2023 was the year that we experimented with it. And 2024 is the year that gen AI has rolled out at scale at companies. And I mean, that's what some executives are saying, but then you look at numbers like you just shared, it's clear a lot of companies are still very early on all of this.
Charlottte Gartenberg: We may be in the early days of generative AI being used in HR, but according to the founder of Girls Who Code and Moms First, the focus should be on tech innovation and mitigating risks. Next up, Reshma Saujani tells us about her organization's new chatbot and why she's optimistic about how AI will be used in the future, both at work and beyond. That's next. WSJ reporter, Chip Cutter, told us we're already seeing chatbots being developed to help employees navigate company policy and answer questions. Reshma Saujani, founder of Girls Who Code, recently helped launch a similar chatbot to do just that. Her goal was to reach further than just one company. Paidleave.ai, a chatbot designed to help people parse the patchwork of laws affecting paid family leave in New York State, went online in December of last year. The tool was launched by Saujani's other organization, Moms First, in partnership with Novy and Craig Newmark Philanthropies. When she first came up with the idea, Saujani says she contacted OpenAI founder and CEO, Sam Altman. She says he put her in touch with the team behind ChatGPT, who provided early technical advice and support. They connected her with Novy, a developer that is part of the OpenAI Service Alliance, who developed Paidleave.ai on OpenAI tools. The goal of Paidleave.ai, to answer questions about leave policies in plain language. I began by asking Reshma Saujani to take me through how the website works.
Reshma Saujani: So let's say you're pregnant and you go to Paidleave.ai. And what it's going to tell you is, am I eligible? How much money am I eligible for? And it's going to give you an action plan. When you go to the government website, they don't really do that for you. And one of the things I love about the site is, it gives you a bunch of easy prompts, because maybe you don't know what you're supposed to ask for. It doesn't just answer your questions, it will draft you an email to HR if you need one. It will let you email yourself an action plan or a checklist. You can save it and send it to yourself and continue to work on it.
Charlottte Gartenberg: So do you see this as a model for other kinds of chatbots being used in HR?
Reshma Saujani: It's very interesting because you can tell on Paidleave.ai, where's the traffic coming from? Most of the traffic's coming from LinkedIn. And if you look and see, it's literally HR professionals sending it to one another.
Charlottte Gartenberg: Well, I find it super interesting that you've talked to HR people that are like, yeah, bring on the chatbots.
Reshma Saujani: We've been operating in a state of fear that we need to move through. Now that doesn't mean that we should be like, ah, there's no risks, it's all good. But that means though that we should be mitigating the risks, but still thinking about the innovation.
Charlottte Gartenberg: Right. So let's look the risks in the face here. What are the risks of creating an HR chatbot?
Reshma Saujani: I mean, I think one of them is that it's going to replace workers. And so I think that there's a lot of fear that a tool like this will make it like you don't really need a human to help you navigate this process, 'cause now you've got a chatbot to do it. I think the second thing is it will give you the wrong answers, which is terrifying, when it comes to these life momentous events that you really need to know, how much money am I going to get or what do I need to save for? And so these risks, or these fears, are real and people are really feeling them. And it's funny, when I started this project, the first question I asked was, what are the hallucinations, do I have to hire a human? You know what I mean? To make sure that they're looking through all the answers. And as I dug in, I learned so much about how this particular use case of paid leave doesn't have the same safety risks because the language model is so limited.
Charlottte Gartenberg: This particular large language model or LLM helps mitigate potential hallucinations. So what data went into the training of this chatbot?
Reshma Saujani: Our AI is trained on the New York State paid leave law. Unlike ChatGPT, it's not pulling from the internet and collecting data from a bunch of different sources. That's when AI bots are encouraged to actually make things up.
Charlottte Gartenberg: Did you weigh hallucination against how much information you could possibly make available?
Reshma Saujani: Absolutely. Listen, I mean, like I said, this is one of the first use cases of doing this. So I did not want to set us up to fail. And so, what's interesting is because there is, again, a limited LLM on this, there's less safety and risk issues.
Charlottte Gartenberg: So with the large language model you're training it on, it's limited to New York state law that helps get rid of some or get around some of the possible "hallucinations"?
Reshma Saujani: Correct.
Charlottte Gartenberg: Yeah. But then we do have the risks of like, is this going to replace humans, the ones that you mentioned? How do we deal with those things?
Reshma Saujani: I don't know. I don't know. I mean, it's so funny. Again, remember, I spend my life teaching girls to code, and I've had moments where like, wait, should they still learn how to code? Because there's this perception that the skills that you're really going to need in this artificial intelligence moment is creativity. You're going to actually need to lean more into the liberal arts because the bots will code for you. So I think that we have to see in many ways how this kind of plays out.
Charlottte Gartenberg: But I still worry about it in the future. I mean, if we see more chatbots replace HR employees, that I won't be able to interact with a human being during a time that's pretty sensitive, it feels potentially dehumanizing. Like I'm just data relating with data.
Reshma Saujani: It might. But what are we going to do? It's here. It's here. We're stuck in this conversation of doomsday, and it's preventing us from really capturing all the opportunities and use cases of AI. Every conversation that we're engaging on in AI, is it good or is it bad? Is it going to destroy us or not? Is it going to replace workers or not? It's here. It is here. And so now the question is, is how do we use the technology in a way to help people, to preserve jobs, to solve COVID, cancer, climate, to do good things. Not having access to paid leave is a major driver of women leaving the workforce or downshifting. And so if I can actually navigate having time off, money in my pocket, the resources I need maybe to hire a care worker, or to get sorted, I'm going to stay in the workforce.
Charlottte Gartenberg: So what's the bigger picture here? Where do you hope Paidleave.ai goes from here?
Reshma Saujani: We're going to expand. The goal now is I want to prove that generative AI can actually increase the uptake of benefits and figure out exactly how to do that. So that means I got to have a handful of governors that are partners so we can work with the Department of Labor to understand, what are the pain points, is our tool solving for those pain points? And now what's the uptake of benefits now that we have this tool?
Charlottte Gartenberg: I'm noticing this trend and how you're talking about things. I think one of the things people think about AI is that it is the solution in itself, that the product, the AI, the chatbot, is the endpoint. You don't seem to be talking about it in that way. You seem to be talking about it as this is the next step in a journey towards other things.
Reshma Saujani: Absolutely. I think it's like a tool, a friend, a helper. I see this from an activist perspective. My sole laser focus is to get paid leave and childcare passed. And so I'm thinking about all the different ways, all tools that we can have to do that. Part of why I think this got so much attention, it was unexpected. It was an unexpected use case of generative AI. And so, I think, as we're looking at what are the use cases of generative AI, go to the unexpected, go to the most vulnerable. The next place I want to go is how am I helping military vets access their benefits? How am I helping people who need food stamps access their benefits? How am I helping people on Medicare access their benefits? Who's the most vulnerable? And how do we build AIs for those communities to help them change the quality of their lives? That's where we need to begin.
Charlottte Gartenberg: The Future of Everything is a production of the Wall Street Journal. This episode was produced by me, Charlotte Gartenberg. Thanks for listening.
Danny Lewis is an audio reporter and co-host for The Wall Street Journal’s Future of Everything podcast, where he covers the latest developments in science and tech and how they will change everyday life.
Before joining the WSJ’s podcast team, Danny was a reporter/producer for WNYC and Gothamist, where he covered arts and culture, local government and environmental news in New York City and New Jersey. He has written for publications including the New York Daily News, Smithsonian Magazine, 99% Invisible and many more. Danny earned his bachelor’s degree from Bard College and his master’s from the Craig Newmark Graduate School for Journalism at CUNY.
In his spare time, Danny is a cartoonist and illustrator, as well as an avid reader of everything from comics to cookbooks. He also enjoys birdwatching, especially for ducks. He lives in Brooklyn, where he was born and raised, alongside his two cats, Soba and Miso.
Alexandra (Alex) Ossola is an audio reporter and producer covering science and tech for The Wall Street Journal’s Future of Everything podcast. Before joining the Journal, Alex ran podcasts and special projects at Quartz and was the managing editor at Futurism. She has written about science for publications such as The Atlantic, National Geographic, Popular Science, and many more. She earned her bachelor’s degree from Hamilton College and her master’s from NYU’s Science, Health and Environmental Reporting Program (SHERP).
She is fluent in Spanish and is learning Korean. In her spare time, Alex likes to travel, bake, read, and birdwatch. She lives in New York City.
