Ep 165: Ethical Leadership for AI Implementation in the Workplace

Episode Categories:

Navigating Ethical AI Implementation

In a digital age shaped by rapid technological advancements, artificial intelligence (AI) has emerged as a transformative force across industries, including the world of work. As businesses seek to harness the potential of AI, ethical considerations surrounding its implementation in the workplace have come to the forefront of discussions on responsible leadership and organizational decision-making.

Ensuring Ethical AI Implementation

The harnessing of AI's capabilities in the workplace demands a nuanced approach from business leaders. Navigating the ethical dimension of AI implementation involves acknowledging its potential impact on the workforce, reframing traditional roles, and addressing the underlying ethical considerations.

Shaping Future Job Roles and Responsibilities

As AI begins to redefine job functions and workflows, leaders are tasked with envisioning the future of work within their organizations. Ethical leadership in AI implementation involves transparency and proactive communication with employees regarding the potential impact of AI on their roles and responsibilities. It is essential to consider the evolving landscape of work and the balance of gains and losses resulting from AI integration in the workplace.

Recognizing Employee Contributions and Consent

In the context of AI implementation, ethical leadership requires recognizing the contributions of employees to the success of AI initiatives. This encompasses acknowledging and attributing the efforts of individuals involved in AI-related projects, ensuring that proper recognition and compensation are extended. Moreover, securing employee consent and transparently delineating the implications of AI for their roles are crucial aspects of ethical leadership in the AI era.

Maintaining Human-Centered AI Development

Central to ethical AI implementation is the recognition of the indispensable role of human expertise and experience in the development and utilization of AI technologies. Amplifying the human element in AI initiatives enables businesses to uphold ethical standards and leverage the complementary nature of AI as an augmentation tool within the workforce.

Embracing Ethical AI Leadership

Embracing ethical leadership in the integration of AI within the workplace fosters an environment of transparency, purposeful collaboration, and proactive engagement with employees. By navigating the ethical dimensions of AI implementation and embracing responsible leadership, businesses can cultivate a culture that leverages AI's potential while prioritizing the well-being and ethical treatment of their workforce.

In the landscape of AI, targeting ethical leadership in workplace AI implementation is pivotal in shaping a future that balances technological advancement with ethical considerations, ensuring that the human element remains at the center of organizational evolution.

Video Insights

Topics Covered in This Episode

1. Impact of GenAI on the Future of Work
2. Ethical Implementation of AI in the Workplace
3.  Impact on Job Displacement and Ethics
4. Business Metrics for Measuring Employee Performance


Podcast Transcript

Jordan Wilson [00:00:18]:

What happens when a company does implement artificial intelligence into the workplace? Right. We always talk a lot about what happens on the front end, but what about when you start to see Vast improvements and efficiencies and and roles completely changing. What happens then? What is the ethical thing to do when you are implementing AI in the workplace. We're gonna be talking about that today and more on everyday AI. Welcome. I'm your host. My name is Jordan Wilson, and everyday AI is for you. It's for all of us.

Jordan Wilson [00:00:55]:

It's so we can all learn what's going on in the world of artificial intelligence and how we can use it to grow our company and to grow our careers. And, hey, our guest today is perfect for that, talking to an AI, Product manager from Indeed. So what what better way to talk about how it impacts both our roles, and our companies? But before we get into that, Thank you for joining us. If you are joining us live, get your questions in. What do you want to know about ethical, implementation in of of artificial intelligence into the workplace. If you're joining us on the podcast, thank you as well. Make sure to check your show notes. We always have, Other resources linked there.

Daily AI news


Jordan Wilson [00:01:32]:

You could even drop us an email. Follow follow me on LinkedIn and connect there and join a future livestream. But, before we get into that, let's go over as we do every day what's going on in the world of AI news. So, politicians are looking to AI for their next jobs. Speaking of jobs. Right? So former speaker of the house Kevin McCarthy is exploring a new a new career in the worlds of in Space and artificial intelligence. So according to an Axios report, he's using his relationships with figures like Elon Musk to bridge the gap between technology and government. So he sees AI as a positive force and wants to keep California and DC to better understand each other, especially when it comes to artificial intelligence.

Jordan Wilson [00:02:13]:

So interesting. Kind of the first, big name politician that we've heard that is looking to transition into the AI space. Next piece of news, Google. Yeah. Sorry, Google, after roasting you the other day on Gemini. Hey. Some some more positive news, but Google Duet is, upgrading its AI assistant. So Google has unveiled upgrades to its AI enabled, assistant, Duet AI, for Google Workspace.

Jordan Wilson [00:02:37]:

So the major development here is bringing Gemini, to Google Duet, which should launch in early 2024 and And improve the tools, capabilities, and understanding and, generating text as well as providing those multimodal capabilities. So that's the big difference between, you know, Gemini versus its previous model, Palm 2, is Gemini is built multimodal, at its at its core versus, you know, Stacking those multimodal capabilities on top. And then Google also announced the general availability of Duet AI for developers and Duet AI in and Security operations. So, yeah, I'd I'd love to hear even from our audience. What are your thoughts on Google, Google's duet? Do you think it's going to compete with Microsoft Copilot? I'd say probably not, but, time will tell here pretty shortly. Last but not least, OpenAI strikes a deal with news publishers. So OpenAI is the owner, obviously, of the popular chat, chatbot chat g p t, and they've struck a deal with European news publishers Axel Bring your SE to bring their content to train their AI systems. So this partnership marks OpenAI's second agreement with a news company and aims to explore the potential of AI powered journalism, which we obviously talked about on the show yesterday with the New York Times, bringing on ahead essentially ahead of AI and creating an AI department in their newsroom.

Jordan Wilson [00:03:53]:

So, this agreement here is worth tens of 1,000,000 of dollars And will reportedly last for 3 years. Alright. That's not all. There's always more news, so make sure if you haven't already, Go to your everyday AI .com. Sign up for that free daily newsletter. Hey. Even speaking of AI sorry. Speaking of OpenAI, there's the, the registrations are back up for ChatGPT plus.

About Madison and AI product management


Jordan Wilson [00:04:15]:

Midjourney is now available on its own website outside of Discord. So We're gonna have all of those things, all of these news pace pieces and a lot more, but that's not what today's about. Today is about talking about ethical. And keyword on ethical. Right? Ethical leadership for AI implementation in the workplace. And I'm extremely excited for today's guest so we can talk about how do we do this ethically. Right. When we bring artificial intelligence, into a workplace, things change.

Jordan Wilson [00:04:42]:

People's roles change. Their jobs change, and it's a lot to, it's a lot to navigate. So, I'm I'm I'm happy to bring on an expert to help us navigate that. So, please help me welcome to the show. Bring her on there. There we go. We have Madison Bonds who is the AI product manager At Indeed. Madison, thank you so much for joining the show.

Madison Mohns [00:05:00]:

Thank you so much for having me. Excited to be here, and, welcome everyone who's watching today. Really excited to hear your thoughts.

Jordan Wilson [00:05:06]:

Alright. And, hey, just real quick, Madison. Just what is what does that even mean? What does an an AI product manager at Indeed do?

Madison Mohns [00:05:15]:

That is a great question. Everyone's out here using AI as a buzzword. So, yeah, I'd love to put some context to that. So, Product management overall is a discipline, at many companies is really focused on bringing forward a strategic vision for a and area of most of the time a digital product. And so product managers can function in a variety of roles. They can work on, product growth. They can focus on, more technical components. They can focus on user facing applications.

Madison Mohns [00:05:47]:

And so there's a lot of different, subsectors, I would say, of what product management entails in terms of, you know, what what does your day to day look like? It it definitely changes all the time. So specifically, AI product management is looking at how can we leverage different types of, artificial intelligence tools In order to, optimize for experiences for users for our site or internally for our own company to move, and progress forward in the most, you know, high velocity, most efficient, and, hopefully, the most broad reaching way to downstream impact, all of the people that are using our products. So, yeah, you can kind of think of that role as, you know, looking internally and figuring out What is the best path forward, but also there are a lot of AI product managers that focus on kind of this more external facing implementation as well.

How Indeed is using AI


Jordan Wilson [00:06:40]:

So, So okay. So now we know a little bit about what an AI product manager at Indeed does, but I think actually it it's it's interesting to talk and to, go back and forth a little bit on what Indeed is actually doing in the space. Right? Because when we talk about shaping the future of work, a lot of people don't know it might start at that place where you're applying for the job or, You you know, applying for your next position. So, Madison, can you talk a little broadly just about how, a large company like Indeed Is even help, you know you know, specifically using AI to create better connections, right, between job seekers and employers And also, you know, how they may even be, you know, impacting the future of work by making those hopefully, you know, better and more personalized connections.

Madison Mohns [00:07:24]:

Totally. Yeah. HR tech and AI is a huge hot topic. I'm sure with the e a the AI, act coming out of the EU, a lot of people Our thinking about this top of mind is one of the categories that is in high risk, and, many, many companies, including Indeed, have been using AI, in their products for years. And, one of the reasons is is that being an HR, person either formally or informally at your company is very hard. When you're sourcing candidates, You're trying to find the right person in this pool of millions and millions of people. And especially for small businesses, you might not Have the time to kind of look through all of those possible candidates that are coming in, reach out to them, and so on and so forth. So, Indeed is really looking at Bridging those connections between employers regardless of how much time they have to contribute to the HR discipline, and candidates that might be a good fit for their next role.

Madison Mohns [00:08:22]:

So we use a lot of machine learning techniques in order to bridge that gap between those two, sides of the marketplace. And so, operationally, like, all of the recommendations that you see on our website, how those recommendations are ordered, and so on and so forth. They're all powered by different machine learning algorithms that intake tons and tons of data that we have about our users, Whether those be our, you know, employers and the jobs that they're posting, as well as the job seekers that are perusing our site and, you know, Demonstrating their preferences for what they're looking for in their next best job.

GenAI and the future of work


Jordan Wilson [00:08:58]:

And, you know, one thing, and we didn't even talk about this beforehand, Madison, but you bring that up, and I'm curious because I do remember, you know, we cover the AI news every day. And, you know, one thing that we've we've we've seen and we've covered before is, you know, All these different reports and better just saying, hey. You know, Gen AI is is coming everywhere. So, you know, even Indeed's AI at work report, you know, showing up here on the screen if listening on the podcast, you know, says, finding that Gen AI will impact almost every job in America. You know, can you, you know, just talk Briefly about how, you know, even just from your personal opinion, how do you think Generate AI is going to change, the future of work? Because I've said it on the show here many times. I said, If you're not already prompting hourly, you probably will be soon in you know, whether that's in, you know, 2 months or 2 years. But how do you see just the the average job, you you know, changing in the future.

Madison Mohns [00:09:51]:

Yeah. That is a really, really great question. There's a there's a lot of research out here about, like, the Placement of work, and I'm happy to get into that more. Many, many tasks are going to be aided by, generative AI systems. I think, that the key difference between and and someone put this in the chat as well. The the the distinction between predictive and analytical AI versus generative AI is that What we're doing here is we're, we're we're generating net new content. Right? It's it's obviously trained off of a large corpus of data that, historical in nature and is pulling from, human insights from across many, many different disciplines. But, the thing that's really great about this is that, it it it shrouds or it gives the illusion of, like, human creativity.

Madison Mohns [00:10:37]:

It's it's coming up with, things that are That can be mimicked, I guess, as what we, we can typically experience in our roles at work. And, I I know personally in my, in my workplace, there are tons and tons of, folks being encouraged to use generative AI in their in their work. I think For me as a product manager, one of the main things I'm responsible is for, is for driving kind of our product vision and strategy forward. And sometimes, it's hard to kind of put words to those things. So, I found myself kind of brain dumping into ChatGPT, like, hey. Here's all the things that are in my head at any given time. I'm working on 10 different projects at once. I'm trying to understand, like, what is that connection between all of these things? And then Once I can kind of, like, pretty much talk through things with, not a real person, but someone that can help me synthesize that information, it can help me, You know, basically make sense of a lot of jargon and, like, random stuff that is coming out of my head And put it into something that is really easy for all the teams that I'm trying to mobilize forward to this strategic vision, and giving Giving some words to, the things that I've been driving kind of in the day to day, but hopefully kind of projecting a more future oriented outlook onto, the work that I'm doing,

Jordan Wilson [00:11:59]:

Yeah.

Madison Mohns [00:12:00]:

With my team.


Implementing AI in the workplace


Jordan Wilson [00:12:01]:

And I I think, Madison, you know, as as we, you know, kinda get back to a little bit more the actual topic of today's show. Right? I think when people think AI implementation, they think gains. Right? They think efficiency. They think, the company's, You know, balance sheet. You you know, they think that AI is just it's it's going to be a win win, for everyone involved, but that might not always be the case. Right? So What are maybe some things that we need to look about look at, especially when it comes to to ethical, implementation of AI in the workplace that maybe things that may not be wins for everyone?

Madison Mohns [00:12:38]:

Totally. Yeah. Let me give you a little bit of context of why I wanted to talk about this today, because this is a I think as a an AI optimist, with a little bit of, you know, cautious optimist, I guess, is what I would call myself. I didn't really realize, when I came into my role that I'm in today, AI was like this exciting new thing. I I I kind of Framed it in the same way that you just framed it as well as many other business leaders framed it. It's a way to make things more efficient and better all around. And, when I was brought on to my team, my main role was to take a primarily, manual operations and transform their existing workflows to be augmented by AI. And so, really, this was, a large strategic focus for our company at the time because we were trying to expand our international influence.

Madison Mohns [00:13:34]:

And in the specific discipline that I was working in, Just called taxonomy. We, we needed to be able to scale very, very rapidly with, as little cost as possible. So for each international market, I'd say, historically, we would be able to have, like, 8 to 10 operations analyst kind of spinning up a market. We wanted to experiment with maybe 1 or 2 people instead of just, like, large, team that we had built out and see kinda what what kinda gains we could get with implementing machine learning. And I was like, that seems like a freaking cool project. I'm really excited to work on this. This was one of, like, my first forays into implementing AI. And when I went and pitched this to my team, they were like, cool.

Madison Mohns [00:14:19]:

But, like, what about us? And I didn't really, I think because of how my specific role was incentivized, like, Mainly, my role as a product manager is to make things faster and increase velocity and make things better overall. But I think in terms of making things better, I wasn't really considering, what was gonna happen to the employees that were training these algorithms that, had the potential To eventually displace their once super multiskilled roles. These are people that are coming out of PhD programs and are, You know, linguists and information architects that have all of this education and training, and I'm reducing them to a labeling task. And I think that is one of the hidden, the hidden things behind Implementing AI is like these models are extremely powerful, but in order for them to be performant, they need to be checked by an expert. And that expert is, you know, in charge of fine tuning those systems. And, in order to do that, They have to dedicate and change a lot of their role to basically train the machine to do their job well. And if we're not intentionally taking time to Help them envision the future of their role. It can be a very, scary and daunting task for those folks.

Madison Mohns [00:15:46]:

And if there's nothing in it for them, why would they wanna go and implement these things? And so that was one of the major things I think on my team When we were incorporating machine learning, it's treating it as more of an augmentation device instead of envisioning Envisioning it as a replacement. We were trying to figure out, like, what are the things that in an ideal day for these, for these analysts, If you had unlimited time, what would you want to spend your time on? And what things can we actually outsource to a machine learning algorithm that, you know, Maybe are core parts of your job, but aren't things that you actually enjoy doing or feel like you are your own specialty and expertise. And so that's one of the things that I think Really what's revealing to me about, like, what is what are the there AI implementation is always seen as good and and Creating gains, but we have to ask the underlying question, like, who wins and who loses in these sit situations?

Jordan Wilson [00:16:47]:

Yeah. And that's such a good point, and I'm excited to to dig down deeper. And if you are just joining us now, make sure to get your questions in now. As a reminder, we have Madison Mons, the AI product manager from Indeed joining us. Madison, you you brought up it's the elephant in the room that no one wants talk about, but I'm always fine talking about it. You know, my thoughts are, you know, implementation of AI, especially across enterprise, will lead a lot of job displacement. People don't wanna talk about it, but I've always said, yes. AI will take a lot of jobs.

Jordan Wilson [00:17:19]:

But ethically, like, Ethics is is the key thing there. Right? So it's like, what will companies do when they find out? Yes. Oh, you know, velocity. Yes. This AI makes us and Faster. It makes us more efficient. Now all of a sudden, you know, instead of people feeling like they have way too much manual work, it might feel like, oh, maybe they don't have enough on their plate or maybe their jobs are kind of like what you said, Madison. Maybe a little more more menial, you know, tagging things, you know, for for models.

Ethical leadership and AI


Jordan Wilson [00:17:46]:

When it comes to the ethics, how should companies be looking at this? Right? Because a lot of companies care about Their shares. Right? Like, like, their shareholders, especially if they're public. They care about profits over people sometimes. How should companies be looking at this? You you know? Because If they do AI implementation correctly, in theory, it is going to completely change how their business operates in a very quick period. How can leaders go about tackling something, you know, that huge, and and that, that carries that much weight? How can they do that?

Madison Mohns [00:18:22]:

Yeah. I think it really boils down to 2 things, which are consent and attribution. And I think the first step here is if you are going to be implementing AI in the workforce, Tell it like it is. You know? Like, we can anticipate that there are gonna be challenges, and I think getting people exciting excited and moving them forward and mobilizing them is great thing, but also making them aware of what actually is gonna happen if this thing works. Right? And having a plan and a path forward. I think it's really important that employees that are going to be impacted in any way, shape, or form, whether that's just a task or their entire role, to understand, like, And be able to visualize from their leadership rather than having to take that responsibility on themselves, To feel like their their leadership is supporting them in their future career path. So I think that's number 1 is really just Being upfront about what this whole thing actually means at that individual Contributor level and what's actually expected of them. And I think the other thing is attribution.

Madison Mohns [00:19:33]:

Like, who actually gets to claim these successes? I think oftentimes in companies, it's people like me who are rewarded and are kind of like the face of this project I get to take all the credit for these huge gains. I get to put on my performance review that I've saved our company over 12,000 working hours by implementing, some sort of automated workflow. And, like, everyone's like, yeah, Madison. You're amazing. But who actually was the ones who were able to actually train that algorithm to be performant enough to actually, save those hours. Right? Like, I think a really good example of this outside of my And, if you guys are interested in this, there's a ton of, literature on this called, it's it's kind of framed around this term called ghost work. So you can think about, like, TikTok, for instance. It's one of the most impressive personalization algorithms that exist today.

Madison Mohns [00:20:28]:

We don't have to see any of the Really brutal or, you know, censored type of content on TikTok because there is a slew of labelers that are, like, outsourced to these 3rd war world countries that are keeping your feeds safe. And so you have to think about the emotional labor of these labelers that are having to watch this extremely traumatizing content in order to keep your Special funny TikTok page of cats and cool grandma videos, you know, safe from having to watch that content yourself. So I think being able to incentivize and and recognize the people that are behind these systems monetarily because, outside of just being like, hurrah, you guys are great at your jobs. You know, these companies are are making tons and tons and tons of money from these gains and efficiencies. And I do think because there's this very extremely detailed, supply chain in which this data gets refined and fine tuned in order to reap these wins and successes. The profits from those types of, you know, huge successes in companies need to actually trickle down to where they originate, in the 1st place. So, yeah, it's kind of a a a mix of, hey. Let's get everyone to consent into this, make it very clear what their path forward is and making sure they're we're upskilling and reskilling them if their roles are going to be displaced or impacted in any way.

Madison Mohns [00:22:00]:

And then 2, just really making sure that we're recognizing this hugely interconnected, group of people that are are really bringing these things to the forefront, in actual through actual attribution and recognition, as well as monetary compensation.

Measuring efficiency with AI


Jordan Wilson [00:22:19]:

I don't know about everyone else joining us live, but this is one of those where I have so many notes. Right? I'm I'm learning so much, along with you. So Let's see if we can continue this learning with with some questions. So, Madison, we actually have a lot, so we'll see if we'll see if we can kinda do a a pseudo rapid fire here. So, Tara's Tara's question great here. So, you know, with AI implementation, what metrics may change when measuring, employee performance. That's a great question because it's like, how do you measure this? Right? What are your thoughts on that, Madison?

Madison Mohns [00:22:49]:

Yeah. That's a really great question. And I think this is a a challenge that a lot of, folks like myself that are working on internal optimization face, Especially when you're working with very small groups of people. So, like, in my team, for instance, I've got a team of about 60 to 70 analysts. So even if I am trying to measure efficiency, you're not gonna get any stat sig types of, metrics that are gonna show like, hey. When we implemented this cooling or this new process or this new workflow. Efficiency gains happened, you know, across the board by this percentage. Right? I think it becomes then tempting to look at individual performance and saying like, okay.

Madison Mohns [00:23:31]:

Before we implemented this specific AI technique, You were producing this amount per day, and now you're producing this amount per day. But I think that really gets besides the point because We then get into kind of this, like, micromanagement type of mode. And so I think in terms of, like, what metrics we need to be changing when measuring employee performance, I we need to be looking a little bit more downstream. So rather than looking at, like, what are my specific employees benefiting from this system? What are the downstream impacts? So one of the things that I work on on my team as the taxonomy team is we're really focused on building out The internal structured data library that describes the world of work. So we've got, an outline of all of the different Skills that are available, different benefits that companies offer, maybe different work settings that are available, different licensure, so on and so forth. And I could look at my team and be like, okay. How many new skill attributes did you produce this month? Right? But, really, what I should be looking at is the performance and the incremental gains and where those pieces of data are actually being used downstream. So the addition of the 6,000 new skills that we identified this month was able to increase our ability to drive connections between employees and job seekers by x percent.

Madison Mohns [00:24:50]:

And so I always encourage people to think about business metrics internally. Obviously, it's it's nice, and it's nice to be able to report out those gains, but in small teams, it's more important that you're empowering people to think about the downstream product impact of the work that they're, actually driving Rather than, like, what are the, like, 1 to 2 types of things that I can slide in there just to meet my quota? So that's how I personally think about it. I like to try to, like, for things as far down the funnel as possible so that I can make sure that my my team is actually thinking not about what is happening at the and Top of funnel and doesn't care about how that actually trickles down. Really encouraging that kinda end to end thinking.

Can AI help with ethics in the workplace


Jordan Wilson [00:25:32]:

Mhmm. That's great there. And if you all miss that one, I mean, talking about how AI gains impact people in the end. Right? I think that's super important. Another another question here from Ben. Ben, thanks for joining us as always. So saying humans already struggle to be ethical. Right? Can AI help? Yeah.

Jordan Wilson [00:25:49]:

Can AI help to implement AI into workplaces? Madison, what's your take?

Madison Mohns [00:25:54]:

Yeah. I do have some hot takes on this.

Jordan Wilson [00:25:57]:

Here's the truth. Here we go. We got to them.

Madison Mohns [00:25:59]:

There's a term called techno solutionism, which is basically saying that technology can solve society's largest problems. I think HR tech is a Perfect example of this. So we've known since the day of the dinosaurs that when humans are hiring people, There are inherent biases. When you're looking at a a person, you're taking in things about their appearance, about how they behave. You can make inferences about where they went to school. All of these things are things that humans naturally are are Assessing when they meet someone in real life. Some of those are, conscious biases. Some of those are unconscious biases.

Madison Mohns [00:26:42]:

And And there's been a lot of, AI products that have been released that claim to be able to take out, human biases and scrub Data sets of what we would call personally identifying information in order to make the most objective hiring decisions. No. I don't think that that actually fully works because there are a lot of implicit things that, a computer can pick up on, that, You know, humans would be able to perpetuate those biases, and machines are also per perpetuating the same biases. A really good example of that, if you Haven't already read about this. This is a couple years back, but, there was a machine learning algorithm that was at Amazon to hire more engineers, and their training data set was a bunch of really successful resumes of, people that already worked at Amazon. And those resumes were primarily from male engineers. And so the algorithm, when it was trying to Find net new candidates based off of that algorithm, started to just kick out all of the women. And it didn't even need to be like, they they had ex taken out all of the names, anything that could be maybe indicative of, like, being a female.

Madison Mohns [00:27:57]:

But at the end of the day, it was still able to pick up on things like if someone went to a woman's college, it would kick them out And so on and so forth. So I think it's really important to notice that, yes, humans are inherently have problems with biases, And we can infuse those either directly and or, machines can kind of pick up on those unconscious biases, from patterns and data that we might not even realize. So I definitely think that there are, there are promises that AI can make more objective decisions, but that really requires that our day our underlying data is not biased. And we know about, concepts like Systemic racism and, you know, racial capitalism. A lot of the documentation that we have in these historical training datasets are reflective of Human biases, and it's very hard to take those things out. So I do think that we'll never really wanna fully automate some of these higher risk types of activities, especially because we don't wanna, you know, unintentionally perpetuate those, and we should always have some sort of human expert in the loop. Hopefully one that is, you know, as free of bias as possible, but really trying to make sure that we have this expert validation That what a model is presenting is, not gonna be harmful to specific groups of people.

Training employees on AI


Jordan Wilson [00:29:18]:

Cecilia with the comment here just hit it on the head saying and Machines learn their biases from the humans that program them. Absolutely. I think we have time for 1 more question, for for from our audience here, Madison. So great one here from Nadia. Nadia, thanks for joining us. So saying, how do you train employees to use AI ethically, and how does implementation go across different generations within the company? That's great because, you you know, business leaders I talk to, especially from smaller or or more medium sized organizations are sometimes hesitant because they're like, okay. Well, what if Employees just use AI to do as little work as possible. And what if I can't measure it, or what if we can't implement it, you know, correctly or ethically? So What's your take on that and how you can actually train employees to use it ethically?

Madison Mohns [00:30:02]:

Yeah. That's a really great question, and I I love the generational take here. I am a Gen zer. Whether I'm proud of that or not is debatable. I'm, like, cussed between millennial and Gen z, and so I am one of the folks that grew up with technology. I was not an iPad baby, so shout out to me for that. But I, I I grew up with the Internet, and so I'm used to things and and things that I think for certain generations, where things that were adopted later in their lives, it can be very scary. And I I've definitely seen that, impact in my own team of people that are familiar with, technology, and I think there's a lot of, like, Obscurity that happens with technology.

Madison Mohns [00:30:48]:

Like, people wanna make it sound cool and fancy. Machine learning is really not That, it's not it it is a black box to a certain extent, but it is controllable also. And I think Being having more education around what machine learning and other types of AI techniques do and what they're trained off of, what they can do, what they can't do. Having that transparency can make things a lot more exciting and a lot less Scary. And I think that in terms of training employees to use AI in their roles ethically, you you really need to be thinking about What like, is having someone use ChatGPT to respond to emails a bad thing? Like, That's debatable. Right? Like, maybe if you're a client facing person, maybe that isn't something we wanna do because we are putting our genuine thought and to the clients that we're working with. But if this is my 15th email of the day of something that probably I could have Solved in a a completely different way. Maybe I wanna make those tasks, easier for myself, and now I have more time to do the task that I'm really actually assigned to do, which is not responding to emails.

Madison Mohns [00:32:04]:

So I think you really need to be thinking about, like, what are your core tasks? What are your core responsibilities? Finding ways to Lessen the distractions of the other types of things that happen at work so you can focus on those things that you're really good at and that your that your employees really want to be working on, And then finding ways to augment those activities, with AI when when those activities aren't necessarily beneficial to the business. So that's kind of how I would think about it. I do think that, like, everyone's gonna have their own opinions on, like, what is ethical use and what is not, and having guidelines around that is, helpful for sharing, but I do think that it is, it's a tough a tough line to cross for sure.

Madison's final takeaway


Jordan Wilson [00:32:47]:

Madison, today's episode, we've covered way more than I thought we could possibly cover in in 30 minutes. But, You know, what would as we kind of wrap this episode up, put the holiday bow on it, what's kind of your one biggest takeaway, you know, for business leaders out there when it comes especially ethical leadership and implementing AI in the workplace. What's the one takeaway? What's the one thing that is, nonnegotiable that they that they need to do in order to ethically implement AI.

Madison Mohns [00:33:17]:

Yep. I think a lot of times mandates for AI implementation happen in senior leadership, Spaces as well as from, boards. A lot of those people don't engage with the folks that are actually doing the work. So I'd encourage you, if you are one of those leaders, go talk to the employees that are gonna be impacted by the implementation of AI. By no means do you need to stop it. You need to be conscientious of the humans that are involved in the process of AI implementation And the impact that it has to their their careers and their livelihoods. Obviously, people that have the type of influence to be able to Spread these innovative ideas across the company already in positions of power, and so we wanna make sure that we are Allocating, time and attention to the folks that are gonna be doing that work, understanding their concerns and their fears, And carving out intentional space for them to have a place in your company's future, whether that be in their existing role Or whether that mean they need to kind of do what it takes to kind of improve the velocity of the area that they're in and then, use their skill sets in a different capacity. So that would be my main thing.

Madison Mohns [00:34:39]:

Just keep the humans in the process. There's There's tons of research on having experts in the loop, and we need to also care about what happens to those experts when their expertise can be Largely mimicked by the technologies that are evolving so rapidly.

Jordan Wilson [00:34:57]:

You know, we always talk and and think, and strategize around ethical implementation. And I think, Madison, what what you just, were able to walk us through today and talk us through Was fantastic. So thank you, so very much for coming on the Everyday AI Show, and and telling us and showing us the path on and Ethical AI implementation. Thank you so much for joining us.

Madison Mohns [00:35:20]:

Thank you.

Jordan Wilson [00:35:21]:

Alright. Hey. As a reminder, this was a lot. If you're anything like me, I have more notes than I can even look at right now. I'm literally typing live as as Madison is dishing out all of her expertise on the topic. So if you haven't already, make sure to go to your everyday ai.com. Sign up for that free daily newsletter. We're gonna be putting all those notes.

Jordan Wilson [00:35:40]:

Yes. Me as a human, I'm gonna get off this call. I'm gonna type up a newsletter so you can read and you can learn. We're gonna share other resources as well as other news pieces and what's going on around the web in the world of artificial intelligence. So thank you so much for joining us, and we hope to see you back tomorrow and every day for more everyday AI. Thanks y'all.

Gain Extra Insights With Our Newsletter

Sign up for our newsletter to get more in-depth content on AI