The AI Values Test: Why Workers Now Judge Employers by How They Use AI, and How Leaders Can Make Responsible Adoption Part of the Employment Deal

[00:00:00] Right now, over half the workforce, like 57% to be exact, is secretly using AI to do their jobs. And they are actively lying to their bosses about it. Which is wild. Right? And it's not because they're lazy or, you know, trying to cheat the company. It's because they are completely terrified of what happens if corporate finds out they've just made their own roles significantly faster.

[00:00:24] So, welcome to the deep dive. Today we're looking at this massive structural shift happening across the corporate landscape. I mean, AI is no longer just this shiny productivity tool. No, not at all. It has literally become a litmus test for an employer's underlying character. And that is a fascinating pivot in how we evaluate companies, right? Because for the past three years, the whole corporate mandate was entirely focused on velocity. You know, just how fast can we deploy large language models?

[00:00:52] Yeah. How quickly can we automate basic workflows? Exactly. The biggest perceived risk in the boardroom was simply falling behind competitors. But that initial frenzy has matured and the workforce is now asking a completely different set of questions. And to figure out exactly what those questions are, we are grounding today's conversation in a brilliant October 2026 article by Dr. Jonathan H. Westover. It's from the Human Capital Leadership Review. Great piece.

[00:01:19] Yeah. It's called the AI Values Test. And whether you are actively job hunting, managing a team right now, or just, you know, trying to navigate your current role without stepping on a landmine, our mission today is to help you decode what a company's AI choices really signal about how they value you. Because the signals are definitely there. They are. We want to help you spot the difference between cheap corporate talk and genuine structural commitment.

[00:01:45] Because the data clearly shows that employees are no longer just passively accepting whatever technology is handed to them. There's this October 2026 GP World at Work report, and they surveyed thousands of global workers. They uncovered a staggering metric. Let's hear it. 58% of employees stated they would outright turn down a job offer if a company's AI values did not align with their own.

[00:02:09] And 92% of executives actually acknowledged their AI strategy impacts talent attraction. I look at that data and, okay, let's unpack this. Because it is really easy for a traditional manager to hear that and just dismiss it. Right. They think it's just resistance. Exactly. They might assume, oh, people are just Luddites. They are scared of the algorithm taking their jobs. But that is entirely the wrong takeaway. That same study shows 73% of job seekers actively want their next employer to embrace AI. Yeah, they want the tools.

[00:02:40] They aren't anti-technology at all. They are anti-exploitation. I mean, they want to use the tools, but they want to know who actually benefits from the efficiency they create. It comes down to this concept in labor economics known as signaling theory. Right. Essentially, in environments where information is asymmetrical, meaning the company knows its long-term plans but the employee doesn't, people rely on observable signals to infer the truth. That makes total sense.

[00:03:04] Right. So if a CEO announces that a new AI tool will streamline operations but simultaneously freezes hiring or quietly adds AI output metrics to quarterly performance reviews... The workforce is paying attention to that. Oh, absolutely. They read those signals. They use them to piece together what Westover calls the AI employment deal. Which is essentially an update to the psychological contract, right? The unwritten rules of the workplace. Precisely that.

[00:03:29] The psychological contract is the set of unwritten mutual expectations between an employer and an employee. When you fundamentally alter how performance is measured or, you know, how job security is defined by injecting automation into the mix without recalibrating that contract... You break the trust. Yeah. Employees experience a profound breach of trust. They start evaluating the company based on enacted values rather than stated values.

[00:03:56] That distinction is so critical for anyone trying to survive in corporate right now. I mean, stated values are the glossy codes of conduct. They're the bullet points in the employee handbook, right? The carefully crafted press releases about empowering the workforce. Yes. But enacted values are the paper trail of actual decisions. When a process gets 30% faster because of an algorithm, who captures that 30%? Does the employee get to leave at 3 p.m.? Or do they just get 30% more work piled on them for the exact same salary?

[00:04:24] And enacted values are impossible to fake over the long term. I mean, you can write the most empathetic memo in the world. But if your middle managers are strictly promoting the people who work 60-hour weeks while heavily leveraging AI... The gig is up. Exactly. The true values of the organization are exposed. It reminds me of the classic unlimited PTO trap. Oh, yeah. Perfect example. Right. On paper, as a stated value, unlimited paid time off looks like this incredible progressive perk.

[00:04:53] We trust you. You are adults. Take all the time you need to recharge. But the enacted value... It's a whole different story. Often it's a psychological mechanism designed to extract more labor. Because the culture praises presenteeism, nobody ever feels like they have explicit permission to take a two-week vacation. Plus, from an accounting perspective, the company no longer has to pay out a crude vacation time when someone leaves. The stated policy is total freedom, but the enacted reality is immense pressure. That captures the dynamic perfectly.

[00:05:21] Employees are inspecting the AI employment deal with that exact same skepticism. But the situation we are in right now is even more volatile. How so? Because in many organizations, the enacted reality isn't just pressured, it is completely invisible. The gap in communication is staggering. If we look at Gallup's tracking data, 44% of U.S. employees say their organization is currently integrating AI. Okay. But only 22% report having a clear plan actually communicated to them by leadership.

[00:05:50] So you have half the workforce acutely aware that the ground is shifting beneath their feet. But corporate is essentially responding with radio silence regarding the actual rules of engagement. Yeah, total vacuum. And in a vacuum of information, human nature takes over. People immediately assume the worst case scenario. We see this in the Pew Research data, where 32% of workers are carrying around this diffuse, lingering anxiety that there will be fewer job opportunities for them personally.

[00:06:16] And that diffuse anxiety manifests in behavior that actively damages the business. This brings us right back to that 57% statistic you mentioned at the top of the show. The people hiding their AI use. Right. A joint study by the University of Melbourne and KPMG investigated how employees actually behave when they are uncertain about a company's AI strategy. They found that 66% of employees are relying on AI output without even verifying its accuracy. Oh, wow. That's dangerous. It is.

[00:06:46] And 56% are consequently making errors in their core work. But the most alarming metric for any executive should be that 57% are secretly hiding their AI use entirely. They're just passing the automated work off as their own manual labor. Okay. But are workers just being paranoid Luddites here? Or is hiding their AI use a completely rational survival tactic when corporate is silent?

[00:07:11] I mean, think about changing the rules of a competitive video game halfway through a tournament without telling anyone. Right. They adapt to survive. Exactly. If a player finds a cheat code that lets them survive the new mechanics, they aren't going to broadcast it to the server administrators. They are going to quietly use it to stay alive. If I figure out how to automate half my weekly reporting by Wednesday afternoon and corporate hasn't given me any assurance that my role is safe. Why tell them? Right. Why on earth would I volunteer that information? I'd just be trading my own replacement.

[00:07:41] What's fascinating here is that concealment isn't a failure of employee morality. It is a perfectly rational economic response to corporate silence. The tragedy for the organization is that AI only generates compound enterprise value when workflows are entirely redesigned around it. You need them to share what works. Yes. You need employees to actively flag errors, share prompt engineering tactics, and safely experiment. You cannot build a high-quality AI-enabled infrastructure on the backs of employees who are actively sneaking around.

[00:08:11] The company ultimately pays twice. They completely lose the trust of their talent pool, and the data they get back is garbage. Which means the obvious solution to shadow AI is for leaders to actually communicate the plan. But jumping into the communication deep end is incredibly risky. Because how you communicate that plan can either establish trust or trigger a total meltdown. Oh, absolutely.

[00:08:34] The source material highlights several real-world memos that show just how badly stated deals can go wrong and, you know, how they occasionally go right. The contrast in executive communication styles over the past few years is a masterclass in risk management. A prime example of a misstep is Duolingo in early 2025. Right. Their CEO, Luis von Ahn, sent out an internal memo in April 2025 declaring a massive internal shift to an AI-first operating model.

[00:09:03] Which sounded very aggressive. It did. He explicitly detailed how the company would stop using human contractors for translation and content generation that AI could handle, while simultaneously placing strict limitations on any new human headcount. And inevitably, that internal memo leaked to the press. Of course it did. And the public and industry backlash was immense. It was universally interpreted as a callous, automation-driven purge of human talent.

[00:09:31] But the wild part is, months later in August 2025, he had to go on a media tour to clarify the memo. He explained that the context missing from the leaked document was that Duolingo never had any intention of laying off their full-time employees. They were just cutting contractors. Yeah. They were protecting their core staff by reducing external contractor spend. The underlying business strategy was actually quite protective of their internal culture.

[00:09:56] But by dropping the headline without the surrounding context, they invited the most hostile interpretation possible. Because it forces the workforce to fill in the blanks with their own anxieties. Now you contrast that specific failure of context with the brutal candor of Amazon CEO Andy Jassy. In his June 2025 memo, he didn't sugarcoat anything. Not at all. He explicitly warned the organization that AI integration would unequivocally reduce the company's total corporate headcount over the coming years.

[00:10:25] Which is a terrifying memo to receive on a Tuesday morning. Definitely. But there is a strange utility to it. It completely cures the ambiguity that causes shadow AI. You know exactly where you stand. But candor about headcount reductions without any attached commitments to transition support or retraining just spikes that 32% anxiety metric we discussed earlier. It clears the air. But it poisons the soil. Exactly.

[00:10:52] Then you have leaders attempting to fundamentally rewrite the baseline of performance, like Shopify CEO Toby Lutka in April 2025. His memo didn't just encourage AI use. It mandated it. A total mandate. Yeah. He tied reflexive AI use directly to performance evaluations and peer reviews. He set a new operational rule, basically saying before any team could request additional budget or human resources, they had to tangibly prove they had exhausted all AI capabilities to solve the problem first.

[00:11:22] And here's where it gets really interesting. Because a mandate like Shopify's completely removes the ambiguity. It forcefully merges AI into the employment deal. But you can't just unilaterally rewrite the fundamental contract of a relationship and expect everyone to seamlessly adapt. No, you definitely can't. It's like being in a marriage where one partner suddenly changes the financial rules of the household without telling the other and just expects them to go along with it.

[00:11:48] A mandate only works if the enacted reality includes the budget and time to actually train people on how to meet that new baseline. Let's look at how these enacted deals play out when the rubber actually meets the road. Yeah, because the sources provide a stark contrast between two massive global companies. We've got Accenture and IKEA. Right. So in September 2025, Accenture announced a staggering $865 million restructuring program. Huge.

[00:12:17] Their executive leadership took a highly calculated hardline approach to the AI transition. They announced that upskilling was their primary strategy, but they drew a very firm line in the sand. Staff who could not be retrained and redeployed on the rapid timeline the business required would be exited from the company. Wow. The enacted reality was binary. You are either reskillable or not reskillable. It is a ruthless execution of the AI employment deal, but again, perfectly transparent.

[00:12:45] But my absolute favorite example from the research is Inca Group, the parent company that operates IKEA. Such a great case study. Over a two-year period, they deployed an AI customer service bot named Billy. And Billy was incredibly effective. So effective that it eventually handled 47% of all routine customer queries across their global call centers. And if you are looking at this through the lens of traditional management consulting, that 47% reduction in call volume is a golden opportunity for a massive corporate layoff.

[00:13:15] You harvest the cost savings and boost the quarterly margin. That is the standard playbook. Yep. But IKEA rejected that playbook entirely. Instead of liquidating those jobs, they took 8,500 call center workers and initiated a massive logistical pivot to retrain them as remote interior design advisors. Brilliant. They took the thousands of hours saved by the AI bot and reinvested them into creating a brand new, higher value, human-to-human service.

[00:13:43] They transitioned these workers from handling angry complaints about missing screws to actually designing living rooms for customers over video calls. Which is incredible. And the financial result is staggering. Those newly minted interior design advisors generated 1.3 billion euros in remote sales in a single financial year. That is the zenith of an enacted value. They didn't just issue a press release about valuing their people.

[00:14:09] They built the physical infrastructure to transition a massive block of employees from a perceived cost center into a highly lucrative profit center. And it highlights a structural necessity that most organizations ignore. Which is? Top-down memos, even well-written ones, are insufficient. To build long-term trust and avoid catastrophic operational mistakes, the workforce requires a systemic voice in how these tools are deployed.

[00:14:34] And the power of worker voice was the defining narrative of the entertainment industry strikes. In their 2023 and 2026 agreements, the Writers Guild of America went to war over the definition of AI output. They really did. They successfully negotiated rigid stipulations ensuring that AI-generated text cannot legally be classified as literary material. Ensuring it cannot be used to undermine a human writer's credit, compensation, or separated rights.

[00:15:03] Now, regardless of where anyone stands politically on labor unions, we're just reporting what the sources highlight here, that agreement is a huge milestone. Right. It is a pristine, objective example of how you take that diffuse existential anxiety we saw in the Pew data and convert it into highly specific, enforceable, contractual rules. We saw a very similar dynamic unfold in December 2023 when Microsoft and the AFL-CIO formed an unprecedented partnership.

[00:15:31] Oh, right. The frontline worker integration. Exactly. The explicit goal was to integrate the perspectives of frontline workers directly into the developmental lifecycle of AI technologies, rather than just imposing the finished tech on them after the fact. And regulators are clearly running out of patience with companies that refuse to establish these frameworks voluntarily. We are seeing hard legal pressure now.

[00:15:53] Illinois amended its Human Rights Act, effective January 1, 2026, to legally mandate that employers explicitly disclose when AI is being utilized in employment decisions, from hiring screening to promotion evaluations. Because the legal system is finally recognizing that you cannot have a proprietary black box algorithm determining an individual's livelihood without mechanisms for transparency and human override.

[00:16:19] Okay. But let me play devil's advocate here and bring this back to the brutal reality of the market. Doesn't forming advisory councils and negotiating with workers just kill the speed of innovation? That's the common fear. Right. Because all of that takes months, if not years. If you are a CEO and your direct competitor is currently slashing 30% of their operational overhead by deploying AI indiscriminately, you are not going to sit around forming an exploratory committee on worker feelings.

[00:16:47] You are going to panic, cut costs to survive, and try to win the market share today. If a company slows down to listen to everyone, won't they lose the AI race? If we connect this to the bigger picture, the economic research strongly suggests that the move fast and break things approach with AI is often a rapid sprint into a dead end. Really? Yeah. Economists Daron Asimoglu and Simon Johnson have published extensive research demonstrating that incorporating worker voice doesn't just placate employees.

[00:17:16] It fundamentally steers the trajectory of technological change. How so? When workers have input, the technology is designed to complement their specialized human capabilities rather than clumsily attempting to replace them. And complementarity is what protects the long-term quality of the product. The ultimate cautionary tale of ignoring this is Klarna. Oh man, the Klarna timeline is incredible to look back on.

[00:17:40] In February 2024, the fintech company proudly announced to the world that their new AI assistant was performing the work of 700 full-time human agents. It was successfully handling 75% of their total customer chats. Huge cost savings on paper. Exactly. Following this, they aggressively froze hiring and naturally reduced their headcount. They optimized entirely for speed and immediate cost reduction, completely ignoring the structural voice of their customer service teams.

[00:18:09] But they hit a massive wall of reality. By May 2025, just over a year later, Klarna's CEO had to give an interview to Bloomberg admitting that they were actively scrambling to hire human customer service staff again. Yeah. Total reversal. Because you can automate the simple stuff. But when complex edge case problems arise, customers get trapped in endless algorithmic loops. The quality of their customer support had absolutely plummeted and customer frustration was spiking.

[00:18:36] They harvested the short-term productivity gains, profoundly alienated their remaining workforce, degraded their core service quality, and then had to fund a massive, highly disruptive reversal just to get back to baseline. Which proves the point. Establishing worker voice and negotiating the AI employment deal isn't a philanthropic exercise in making people feel warm and fuzzy. It is a structural safeguard against executing catastrophic business errors that destroy enterprise value.

[00:19:04] So what does this all mean for you listening right now? We've looked at the data, the anxiety driving shadow AI, the absolute chaos of top-down memos, and the massive unforced errors companies make when they chase speed over quality. If you are evaluating a company today, how do you identify an organization that is actually building a trustworthy, resilient AI culture? The research distills this down to three tangible long-term capabilities you should look for in any employer.

[00:19:34] First, the organization makes their AI values empirically measurable. What does that look like in practice? They aren't just looking at log-in metrics to see how many people are using the tool. They are running pulse surveys to track whether employees actually trust the tool. They are measuring how many hours were saved and critically tracking how many of those employees were redeployed into higher-value roles versus how many were quietly managed out. They track the psychological safety, not just the software adoption?

[00:20:03] Exactly. The second capability is that they formally integrate their AI stance into their employer value proposition, their EVP. During the recruitment phase, they are explicitly upfront about how the organization utilizes AI, the specific training pipelines they offer to keep your skills relevant, and the guardrails they have in place to protect your role as the technology scales. And the third capability, which honestly feels like the linchpin to this entire transition, is how heavily they equip their middle managers. Oh, this is so crucial.

[00:20:33] You can have a visionary CEO who writes a flawless, empathetic memo about AI complementarity, but if your direct supervisor doesn't understand the strategy, the entire culture falls apart at the team level. The global data highlighted that only 47% of employees have received any formal AI training from their organization. If middle managers lack foundational AI literacy, they literally do not have the vocabulary to evaluate AI-assisted work, let alone facilitate a nuanced, honest conversation about job security with their direct reports.

[00:21:02] As an employee, you don't experience the CEO's stated values. You experience the enacted values of the most anxious or careless manager you happen to report to. Middle management is the crucible. It is where the glossy stated values either materialize into enacted reality or where they go to die. Which leaves us with a totally fascinating and honestly slightly terrifying paradox to think about as we wrap up. I know where you're going with this.

[00:21:30] As organizations get more sophisticated, they are increasingly building AI into their employer value proposition. And simultaneously, they are using highly advanced automated systems to screen initial candidates for culture fit. It's wild. We are rapidly approaching a truly bizarre future for the job bargain. Think about it. How will you, as a human job seeker, manage to accurately judge a prospective employer's enacted AI values?

[00:21:58] Actually, I shouldn't pause there. How will you judge their enacted AI values when the very entity interviewing you, screening your resume, and selling you on those cultural values is an AI that has been programmed to perfectly mimic the company's stated values? It creates a fascinating closed-loop dilemma. You are actively trying to inspect the foundation of the house, but the house itself is the one generating the structural inspection report.

[00:22:27] It really makes you wonder how long you can trust the fresh coat of paint before you have to start tearing into the drywall yourself. Thank you for taking this deep dive with us today. Keep your eyes open. Watch the paper trail of what they do, not just the press releases of what they say. And we will catch you on the next one. Watch the paper trail of what they do.