Adam Larson: 00:05

Welcome back to Count Me In. I'm Adam Larson, today I'm joined by Doctor. Zahida Daruwala, an academic, researcher, and author for a timely conversation about the real cost of rapid AI adoption in business. We know automation brings obvious changes like layoffs, but Dr. Daruwala dives deeper revealing the hidden impacts from shaken employee morale and lost institutional knowledge to the risks of organizations becoming indistinguishable.

Adam Larson: 00:32

We talk about leadership challenges of balancing shareholder demands with employee dignity and why accountability for AI can never just be handed off to an algorithm or a vendor. So, if your organization is figuring out how to use AI responsibly and keep its people front and center, this episode is for you. And as you listen, ask yourself, is your company using AI to make people more valuable or just simply less expensive? Let's get started. Well, Zahida, thank you so much for coming on the podcast.

Adam Larson: 01:08

I'm really excited to have you here. And we're going to be talking a bit about the AI layoff loophole, which is something that when you and I first chatted about, I was like, Woah, I've never talked about this before. Let's really get into it. And maybe we could talk about walking back to where this idea started for you. When did you first notice that that conversation around AI in the workplace has shifted from curiosity and experimentation to something that felt more like a race to cut costs?

Adam Larson: 01:33

Because I think that's what everybody's really worried about at this point.

Zahida Daruwala: 01:35

Okay. Yeah. So let me tell you one thing. It wasn't like something like one fine morning I got up and that, hey. Is this AI thing really troubling us or no?

Zahida Daruwala: 01:45

No. It it wasn't that sort of experience at all. It was just that it's happened gradually over a long period of time just listening to news and hearing things about what was AI doing and things like that. So let me give you let me start with things like initially when AI was in full flow, people were very curious as to know what can this technology do. Organizations were experimenting.

Zahida Daruwala: 02:11

Employees were learning. But what we see today now more recently, questions are more like instead of asking how can AI make our people more capable, we are asking questions like how many people can AI replace? And that's when I realized that yeah, we had crossed an important line. And when we hear conversations in about decision makers in conferences, about job descriptions and things like that, it's more about measured by the headcount reductions rather than value creation. So that's when I got a feeling as to what is happening in our corporate world related to AI.

Zahida Daruwala: 02:55

Is it being just been treated as a cost cutting tool? So I'll give you one news article which I was just going through and it just hit my thought is that big names like Amazon, Citigroup, Dell, HSBC, Intel, Meta, Oracle, UPS for that matter, when we hear that there are so many layoffs because companies trying to be more efficient with AI tools, I think that's when I started thinking that yes, I mean, what is it that's happening in the corporate world? Yeah, that's how it's got it's got my interest rolling on.

Adam Larson: 03:36

Zahida, thank you so much for coming on the podcast. Really excited to chat with you. And we're gonna be talking a little bit about the AI loophole, the AI layoff loophole. And it's something that when you and I first chatted about it, was like, okay, people aren't really talking about that, but it's something we're starting to hear. And maybe we can start off our conversation by walking back to where did this idea start from you?

Adam Larson: 03:54

When did you first notice that in the conversation? You know, because a lot of times people are when you get to AI, they're talking about it seems like it's a race to cut costs and eventually people are going to start getting laid off because of that. And so what does that look like and, you know, when did you first start seeing that?

Zahida Daruwala: 04:08

So to begin with, actually, I was also quite curious about knowing what companies are doing with AI and how are they implementing that in their, in their day to day environments. So, yes, I was a little curious. And when initially when conversations were being around what AI can do for companies, people were experimenting, employees were learning. More recently, questions started changing. Instead of asking how can AI make our people more capable, I started experiencing or feeling the questions were more around how many people can AI replace.

Zahida Daruwala: 04:48

And that's when I realized that something important is necessary to investigate. So when I start reading about what companies are doing with AI, there are news headings which come up like, so many CEOs have announced layoffs due to efficiencies made by AI. We hear company names like Amazon, Citigroup, Dell, HSBC, Intel, Microsoft, Oracle, UPS. So when you hear such names and layoffs in these companies, it's really interesting. I would say it's really interesting.

Zahida Daruwala: 05:30

So there was one article which actually said that when you compare from 2024 to 2026, we found that the layoffs due to AI had increased eight times more in the current year just three, four years back. And 60% of those layoffs were primarily due to it was basically for those companies that were more than 100,000 employees. That is we are looking at large companies having layoffs. So, yeah, that's when, you know, I started a little bit reading more about what AI is doing in in the corporate world. Yes.

Adam Larson: 06:07

That's really interesting. And and I think all of us have kinda seen those articles and and been worried because, you know, what does it look like for my organization? And and sometimes people just start saying, hey, AI is just a tool. AI is just a tool. It's just a tool in your toolkit.

Adam Larson: 06:20

Don't worry about it. Don't worry about it. But, you know, and maybe we could talk a little bit about that because that framing is not necessarily wrong, but, you know, it was it letting organizations off the hook when it comes to how they're treating people?

Zahida Daruwala: 06:31

So I actually agree. I agree a 100% with you. AI is just a tool, and I think organizations can't use that as an excuse. Like we say, a hammer is just a tool. It can build a home or break a window.

Zahida Daruwala: 06:51

We don't blame the hammer. We ask who decided how to use it. AI is no different. So the technology itself isn't ethical or unethical. It's the decisions that people make about how to use AI and carry out their work responsibly.

Zahida Daruwala: 07:10

That is what matters. So many times we hear that if something goes wrong and, you know, there are some hookups or some, you know, bad actions or something like that, We hear in the news that AI made that decision. But don't we think a second time. Is it actually AI that's made the decision or is it the people that decided to use the AI took that decision wrongly? So let's take an example of a company that replaces an entire customer service team overnight because an AI system can handle tasks more cheaply, probably more cost effectively.

Zahida Daruwala: 07:56

Was that an AI decision? I would say no. That's a leadership decision. That reflects an organization's values, priorities and the definition of success. So the big question today is can we trust the machines?

Zahida Daruwala: 08:12

When the company replaces an entire department with an algorithm overnight just to show good numbers to please shareholders, is it smart business or just a violation of corporate social responsibility? Technologies can make decisions faster than humans, agreed. But who is responsible when something goes wrong? Today the ethical question is not whether an AI can perform a task better or faster. Of course it can do.

Zahida Daruwala: 08:40

We know that. But the real question is how do we use AI in a way that enhances human potential and not just focusing on reducing head count.

Adam Larson: 09:03

So there's something I wanna I wanna pick on that you just talked about. You talked about how, like, a company automates a system overnight, you know, and then in like, what is actually being lost when you automate something overnight? And, you know, what are the other hidden costs, especially for the organization when that happens?

Zahida Daruwala: 09:18

So I think the headlines focus mostly on the visible cost. That is the layoffs which is quite understandable, quite measurable. But what we need to understand is over and above that this is just the beginning of the story. The real losses are often invisible. They don't show up immediately on the balance sheet nor on the income statement as reduced costs.

Zahida Daruwala: 09:44

So let me just tell you something about what are the hidden costs that we should consider the corporate world. So the first thing I would like to highlight here is about employee morale. So imagine watching colleagues disappear overnight because their work has been automated. Even if you keep your job, you're left wondering, am I next? That uncertainty changes behavior, changes behavior of existing employees.

Zahida Daruwala: 10:14

People become more cautious. They are less willing to take risks and they are less likely to contribute to bold ideas. Innovation doesn't thrive in such an atmosphere of fear. It only thrives when people are feeling free and secure enough to experiment and challenge conventional thinking. So yes, the first hidden cost which might not reflect itself anywhere is the employee morale.

Zahida Daruwala: 10:42

Second thing that I would like to also emphasize here is the loss of institutional memory. Let me explain what do I mean by this, the loss of institutional memory. You see, whenever any experienced employee moves out of the door, he not only carries with him his biodata but also his years of tacit knowledge, those unwritten insights about customers, relationships, judgment, about problem solving, which has never made its way to a manual or a database. So AI can analyze historical data remarkably well but cannot inherit the context, the intuition, the lived experience that people can accumulate over time. So another hidden cost, I would say here, is the loss of institutional memory which people or organizations or leadership may not realize now.

Zahida Daruwala: 11:47

But yes, that's a big loss. That's what I feel. And thirdly, I would also like to add something known as homogenization of work, which I think is more of a long term risk rather than having or showing its immediate colors, as you can say. It cannot be immediately visible. So as we know, AI learns from patterns which are there existing in the data and continues to repeat those patterns.

Zahida Daruwala: 12:22

So that's what I think about is AI does is homogenize work, and there's a risk to that as well.

Adam Larson: 12:30

Yeah. That homogenization is something that I don't think people are talking about enough because, you know, you're you kind of lose that diversity of thought within your organization, which drives innovation because, you know, your experience is different from my experience. The way you see the world is different from the way I see the world. So you'll bring a perspective that I may not have never ever seen before. And that opens my eyes and I say, well, have you thought about it this way?

Adam Larson: 12:51

And then with our two perspectives, we can bring something new that might not have been seen if our two perspectives weren't in there. How do you explain that dynamic to like a leadership team?

Zahida Daruwala: 13:00

Exactly. I agree. I agree to that. Yes. Yes.

Zahida Daruwala: 13:03

So, let me tell you something more about what is this homogenization and how it feels like that's the hidden cost that corporations and leadership is simply ignoring. Okay. So let's say if an organization relies on AI models which are trained similarly on similar data sets, there's a real danger that everyone starts solving problems in the same way Because it's not people who are solving problems, it is actually the AI models that are solving problems. So we may be incredibly efficient at producing now the expected answer. So the expected answer is what AI generates while losing creativity, because creativity only comes when we bring about unexpected results, not the expected ones.

Zahida Daruwala: 14:00

So let's say there are 10 competing companies in the particular industry. They are all trained on the same AI model and they analyze market data. They recommend the best strategy. And the chances are that they'll all receive remarkably similar recommendations. So they may all target the same customer market, they may launch similar products, they may adopt nearly identical pricing strategies.

Zahida Daruwala: 14:27

But you know today competition isn't about doing exactly what everyone else is doing. AI is excellent at finding patterns in the past. But innovation comes from imagining something that has never been done. Let's take an example of our mobile phones. Today we have these smartphones.

Zahida Daruwala: 14:46

So if you had AI analyzed customer data and AI was collecting data about people's preferences and what they wanted, people could only imagine a mobile phone at that time, years passed, with a long battery life. Or say, for example, with improved buttons. Very few would have imagined a mobile with a touchscreen computer. So this breakthrough only happened when somebody imagined that even a touchscreen could be implemented in a mobile design. So, you know, Steel Jobs famously argues that people don't know what they want until they see it, and that's the difference.

Zahida Daruwala: 15:26

AI can tell us what people wanted yesterday, but human creativity imagines what they'll want tomorrow. So, yes, I would say AI is brilliant at recognizing patterns, but humans are brilliant at breaking them. If organizations rely on AI, they'll become increasingly efficient at repeating the past but the future belongs to those who can imagine something different. So this is what I mean by the hidden cost of homogenization. So that's what I meant, which now corporate leaders should be thinking about.

Adam Larson: 16:04

Yeah. And and everything you were saying made me think of like those, those flyers, those online flyers you're seeing now across all the social media and all the flyers look exactly the same. They're using the same icons. They just have the different text and you can tell, hey, everybody use some sort of AI to create this. And it's and they break all the rules that traditional flyers have.

Adam Larson: 16:23

They have way too much text, way too busy, way too much going on. You don't really can't really follow it. And it seems like and if all the organizations are using the same flyers, then you need to do something different to stand out from that. Is that that same kind of philosophy that you were saying where we can't just homogenize it? Everything can't look the same.

Adam Larson: 16:38

Memory can't think the same way because you won't have the innovation that we all need desperately as organizations.

Zahida Daruwala: 16:44

Exactly. Exactly. And without innovation, how would we grow? I mean, how would we create something new for the world, for the people to be happy and more satisfied as times go on? So, yes, AI is good, but how good?

Zahida Daruwala: 16:58

That's what we need to think.

Adam Larson: 17:00

Yeah. So, you know, there's a real tension, you know, in organizations because, you know, especially when you work for a public company, their shareholder expectations, they want that bottom line, they want their return, but there's also employee dignity. There's also that, you know, helping your employees feel like they're they're part of the organization, that they're they're they're giving something back and they're they're part of this greater good and also keeping, you know, their jobs alive because we all need jobs to survive. You know, how you know, from where you sit, know, how do leaders actually kind of navigate that that tension and what version of this doesn't require choosing one over the other? Like, you can have some employees or, you know, we're we're gonna help the we're gonna help the shareholders, but we're gonna get rid of all the employees.

Adam Larson: 17:37

You know, there's even been, like, shareholder calls where the where the guy was like, yep. We got rid of 72,000 people. And you're like, wait. Wait. Wait.

Adam Larson: 17:43

Wait. You just fired 70,000 people. And like, what about how those people they're they're they're not just a bottom line. They're human beings. And how do you fight how do you like kinda balance that tension?

Zahida Daruwala: 17:51

So let me tell you, there is there is actually no tension. This is this seems to be like a false false dilemma. It's not a true dilemma. That's not how how we should think as as corporate leaders or decision makers. So I think, I believe that it's not a choice between shareholders and employees.

Zahida Daruwala: 18:14

And leaders should understand that employees aren't simply a cost to be managed. They are a source of long term value. If you treat people well during technological change, you are not working against shareholder interests. You are actually protecting them. So we see organizations will succeed in an AI era will be those that don't just automate fastest.

Zahida Daruwala: 18:43

So it's not about just automating things the fastest. It's about building trust. Let's say, like for example, we simply can't say that how many roles did AI eliminate and that was one of the success factors. Instead, leaders should ask how many people can we transition into higher value work? That's where real business ethics begins.

Zahida Daruwala: 19:12

Okay. So we say AI will automate repetitive tasks, but then organizations have to invest responsibly in reskilling, upskilling and creating pathways for employees to evolve alongside technology. So workforce transition is more sustainable than workforce reduction. So we should not focus on workforce reduction as a measure of success. So that's what I feel that that's where the challenge or the tension lies in the way that leadership thinks.

Zahida Daruwala: 19:52

And moreover, what I would like to say add here is about how do organizations recruit people. So while recruiting people, also there is a lot of use of AI. For example, in screening CVs, ranking candidates, recommending promotions, evaluating performance. If systems can do this, yes, systems can be efficient. But can they be fair?

Zahida Daruwala: 20:22

They can only be fair as far as the data and the assumptions are behind them. Now historical data can be biased. AI can unintentionally scale that bias across thousands of decisions related to human resource or related to recruitment, related to appraisals. So that's the danger that what I want to highlight. That's what can then damage reputation.

Zahida Daruwala: 20:50

It can manipulate markets and also erode confidence in institutions. So the biggest question here is not the tension between employees and investors or shareholders. It should be about accountability. So you cannot allow an AI system to reject a qualified job applicant and make biased decisions. Who is responsible for that?

Zahida Daruwala: 21:15

Is it the software? Is it the HR manager? So my view is very clear in this. Accountability can never be outsourced to an algorithm. Leadership means, okay, use technologies but you choose how to deploy that.

Zahida Daruwala: 21:32

So we say AI can empower shareholders, empower employees, and then affect shareholder outcomes, but it should be coupled with good ethics.

Adam Larson: 21:42

So I think when we're using AI, there's something that we kind of that's not getting talked about as much where when AI gives wrong information to to customers or shares things it shouldn't or, you know, the hallucinations, you know, maybe without getting too much into the details, you know, what are what what is kind of some of those incidents look at? Like, you know, I know there was an incident where, one of the companies, an ERP, was giving financial information to other, you know, to customers that it wasn't supposed to be. You know, what does that reveal when like the human checkpoints are kind of removed from financial systems in the name of speed? Because you're like, hey, we wanna do this faster, but if we lose those checkpoints, what what's happening?

Zahida Daruwala: 22:20

So see, this matters too much when it comes to financial information. Okay? When it comes to customer information or behavioral or other social information, it may not impact so much. It may just be forgotten after a certain period of time. But when it comes to financial information, I think we need to be more careful when dealing with that.

Zahida Daruwala: 22:41

So let me give you an example of Sage. So Sage is one of the companies that deals with, it's one of the world's leading accounting providers as we know and it had recently introduced an AI powered co pilot which could make the financial workflows more faster and efficient. So what happened was that shortly after the launch and implementing it, there was some glitch reported in the Copilot where customers could see other customer information when they asked for their information. So basically like if one customer would ask the AI app that I would like to retrieve for me this month's invoicing data or something like that. It went on and gave other customers information and some customers contacted Sage and said that hey look here we've got information which is not our account it's somebody else's account.

Zahida Daruwala: 23:40

Now imagine you're one of the customers of Sage and you asked that particular app to provide you that information. Yes, you got it instantly. You got it very fast. But then you also got somebody else's data, somebody else's maybe confidential information, bank accounts, maybe some other you may call it some IDs, ID numbers or some confidential information. Now you as a customer would think that if that information has flowed to you, what are the chances that your information has gone to someone else?

Zahida Daruwala: 24:20

So Sage took it very seriously, yes. Sage, when they came to know from customers that they were getting this, some other information also being fed into their requests, So they immediately tried to find out what is the glitch. They stopped the AI. They went into rectifying it and then came back. And then they said that, oh, that was only affected only a few of the customers.

Zahida Daruwala: 24:43

The question here is not about how many customers did it affect. The question is about even if a small amount of information could have possibly been leaked or misappropriated or misallocated between customers, What are the chances that it may not happen again? So that's the real lesson of how that should be dealt with, when it is confidential, personal, or financial information.

Adam Larson: 25:13

Then I guess in the question, I I question, you know, where who's accountable for that at the end? Because, you know, who owns that outcome? Because is it the employee who approved the system? Is it the company who deployed it? Is it the vendor who built that?

Adam Larson: 25:25

Like, who actually is holds the accountability? Because I feel like when it comes to corporations, everybody kinda pushes accountability to other people. And so I wonder, like, what like, who who's accountable for those things?

Zahida Daruwala: 25:37

Yeah. That that that's a great that's a great question to ask. Actually, everyone should ask if there is any glitch in the AI, who's responsible for it? And of course, that's the biggest question that we are also fighting with today. So let me give you a simple example.

Zahida Daruwala: 25:55

Like for example, we have aircrafts and aircrafts fly on autopilot, right? They fly on autopilot most of the time and most of the journey is smooth and comfortable and accurate. But what if something goes wrong? Can you simply blame the autopilot for doing something wrong? You have to examine the whole ecosystem.

Zahida Daruwala: 26:18

Was the technology designed properly? Did the airline maintain it? Were the pilots adequately trained? Did they intervene when they should have? So when we look at AI and who's responsible for and who's accountable for any glitch or any issue that comes up, We cannot simply say that AI made the decision.

Zahida Daruwala: 26:41

Of course, regulators will not accept that defense. Organizations chose to deploy the system. They selected where to use it and they are responsible to ensure that appropriate controls are in place. So yes, of course, I would first and primarily give the accountability to the organization as a whole. At the same time, it is also the vendors.

Zahida Daruwala: 27:03

It is the people who supply or create those apps and that technology. They also have an obligation to build secure, reliable and transparent systems. And they should be able to fix defects whenever they are identified. So yes, I would say the line of responsibility also falls in to the software vendors. And of course the employees.

Zahida Daruwala: 27:25

The employees who are actually working on behind the AI within organizations cannot be completely removed from the picture. They can't just wash their hands off. If an AI recommended something that looks unreasonable and someone approves it without applying their professional judgment, they also share the responsibility. So today in the AI world, accountability is distributed across everyone involved. We can't pinpoint only one party.

Zahida Daruwala: 28:00

Accountability multiplies with AI. So as an organization employs AI for their use, they should remember that accountability is across the board. AI can assist with decisions, but accountability must remain with humans. So, yes, deploy AI, but multiply accountability. That's what it needs.

Adam Larson: 28:27

Yeah. I no. I appreciate that. Multiply accountability. Like, everybody's responsible.

Adam Larson: 28:31

You know? But when when we think about the accounting and finance team and, we know AI is here. What does it look like? What does responsible AI advocacy look like from inside that team? You know, you're seeing what's happening at an organization level.

Adam Larson: 28:44

What does that advocacy look like from that perspective?

Zahida Daruwala: 28:47

So, see, again, the primary responsibility is the organization as a whole. So it's the decision makers at the They cannot wash their hands off. They cannot pinpoint to the employee. Okay. You did not do that.

Zahida Daruwala: 29:01

And then you're out. That's that's not the right approach. Okay. So yes, it falls down from the top management. They are responsible first and of course, yes, ultimately if there is found a fault with the employee who ignored that particular red flag or approved it without his own judgment, provided that powers were given to him when such an incident happens.

Zahida Daruwala: 29:28

So if you rely more on AI rather than the human so that's from the top, right? So when the top decides to give more power to AI than the employees, that's when the problem occurs. So that's what I wanted to point out here. It's across the board. And of course, there should be responsibility as well as authority.

Zahida Daruwala: 29:48

So you make the employees responsible, but also give them the authority to stop you when they see something wrong.

Adam Larson: 29:56

I there's something you've you've you've told me previously, like, real the company's real code of conduct isn't on their website. It's visible in their budget and how they treat people during especially during a technological shift. So, you know, as we kind of wrap up this conversation, what do you want our listeners to walk away from asking themselves about their own organizations after this conversation?

Zahida Daruwala: 30:15

So I would like my listeners to look first within their own companies and ask one simple question. Are we using AI to make people more valuable or simply less expensive? Again, I'll repeat. You say they can ask themselves when they look around, are they using AI to make people more valuable or simply less expensive? Because the answer to that question reveals far more about an organization's values than their mission statement on their website or in their documentation.

Zahida Daruwala: 30:54

Technology isn't itself ethical or unethical. It's a tool. We said that earlier. It's just a tool. Ethics comes from the choices that leaders make about how the tool is to be used.

Zahida Daruwala: 31:07

So do we use AI to eliminate the repetitive work so that people can focus on creativity, strategy and solving problems? Or are we simply using AI to cut costs? Look at that around within your organizations. And that's what will give you an understanding of how AI is being used within organizations. And I believe that organizations that will thrive in the coming times won't necessarily be the ones with most advanced AI implementation.

Zahida Daruwala: 31:44

They are going to be the one that use AI along with strong human judgment, ethical leadership, coupled with a commitment to developing their people. So I'd like to leave the listeners with one final thought that if we look back ten years ahead, I mean, back from ten years ahead, like if we are ten years ahead and look back again, we'll probably remember those we won't remember the companies that adopted AI first, but we'll remember those companies that adopted AI responsibly. So yes, this is my closing thought that while embracing technology, companies should not lose their humanity. AI with humanity, and that's the tagline that I want to close with.

Adam Larson: 32:41

Zahida, thank you so much for coming on the podcast. It's been amazing listening to your insights, and I really appreciate the perspective you've brought for our audience today. Thank you so, so much again for coming on.

Zahida Daruwala: 32:51

Thank you, Adam, for having me, and I did enjoy talking to you as well.

Announcer: 32:56

This has been Count Me In, IMA's podcast, providing you with the latest perspectives of thought leaders from the accounting and finance profession. If you like what you heard and you'd like to be counted in for more relevant accounting and finance education, visit IMA's website at www.imanet.org.