Abstract: Early evidence on generative artificial intelligence at work suggests that its benefits are real but unevenly shared. Year-two results from the American Job Quality Study show that most employees who use AI report working faster and more creatively, yet only 28% of U.S. employees use AI at least weekly and 54% have never used it in their role. Regular use is concentrated among college graduates, managers, higher earners, and knowledge-intensive industries, the same groups that already hold quality jobs, while about half of all employees report less influence over technology adoption than they would like. This article defines the "AI dividend gap," reviews research on who adopts and who gains from generative AI, and examines organizational and individual consequences. It then synthesizes five evidence-based responses, illustrated with documented examples from entertainment, technology, healthcare, hospitality, retail, and financial services: designing for broad access, closing the use gap, building worker voice into technology decisions, bargaining over how gains and risks are shared, and reinvesting gains in better jobs.
Powered by the WRKdefined Podcast Network.
[00:00:00] I want you to think about your work day for a second. Do you feel like AI has actually made your job better or just kind of more crowded? Oh, that is the perfect question to start with. Right. Because companies are dropping billions on AI right now, mostly, you know, to make their top performers even better. But new data shows those elite workers are getting almost zero benefit. Yeah, essentially none at all.
[00:00:22] Meanwhile, the rookies like the frontline workers, the new hires, they could be getting this massive overnight 40% boost in performance, but they're being systematically denied access. So welcome to the Deep Dive. Glad to be here. Today our mission is to really uncover who is actually benefiting from AI in the workplace and, you know, who's just being left behind to do the busy work.
[00:00:46] Yeah, the way organizations are deploying this stuff is creating this fascinating and, well, kind of alarming economic divide. Definitely. And to figure out why, we've got a great stack of research today. We're looking at an eye-opening 2026 article by Dr. Jonathan H. Westover from the Human Capital Leadership Review. It's called the AI Dividend Gap. Right. And he's really done the heavy lifting for us here, synthesizing these massive brand new data sets.
[00:01:12] Yeah, we're talking the October 2026 American Job Quality Study from Gallup. Some deep dive field research out of MIT and Harvard. And real world corporate case studies. So, okay, let's unpack this. What exactly is the AI Dividend? So think of the AI Dividend as the bundle of tangible benefits these tools give a worker. Like saving time. Exactly. Time saved on those routine mind numbing tasks is the obvious one. But it's also, you know, better output.
[00:01:42] Right. Or being able to quickly brainstorm creative solutions. Right. It gives you the mental bandwidth to actually focus on the strategy part of your job. And the baseline from all this research is that, well, the technology works. It does deliver. But the rewards aren't going to everyone. No, not at all. The crisis here is all about distribution. The dividend is flowing almost entirely to people who already have great privileged jobs.
[00:02:09] So as you're listening to this, I really want you to think about which side of this dividend gap your own job falls on. Pay attention because understanding this is the first step to getting your fair share. Let's look at the actual numbers of who gets to use these tools. Yeah. So that October 2026 Gallup data paints a really divided picture. Right now, 28% of US employees use AI at least weekly, which I mean sounds like a decent chunk.
[00:02:35] It does sound solid. But the flip side is that a staggering 54% of employees have never used it in their role. Like not even once. Wow. Over half. But for that 28% who do use it, it's transformative. Right. I think the data said 63% work faster. Yeah. 63% work faster and over half say it helps them find more creative solutions. But the demographic split beneath those top line numbers is where the gap really hits you.
[00:03:04] Right. Let's get into that. So 40% of college grads use it regularly, but only 17% of workers without a degree do. That's a huge drop off. It gets worse. Look at the corporate ladder. 37% of managers use it compared to just 25% of individual contributors. Huh. And entire industries like healthcare support, manufacturing, retail sales, they essentially have zero access.
[00:03:31] It's like, it feels like AI is this exclusive VIP lounge at the airport. Oh, that's a good way to put it. Right. The people already flying first class get ushered in. They get the free drinks and the fast Wi-Fi while everyone in economy is still stuck at the gate paying six bucks for water. But let me play devil's advocate for a second. Sure.
[00:03:51] AI models are mostly text based, right? They write emails, summarize reports. Isn't it just natural that white collar knowledge workers are the ones using it just because of what their tasks are? I mean, task exposure is definitely a piece of the puzzle. If your whole job is spreadsheets and memos, it's a natural fit. But the research points to something way more systemic. Like what?
[00:04:13] Organizational choices. Those are the real gatekeepers here. The gap exists because of choices about who gets a paid enterprise license or, you know, who gets a secure company device to run the software. Oh, I see. And most importantly, who actually gets explicit permission and paid training time. A retail manager might have a hundred ways AI could help with scheduling, but if corporate doesn't give them the tool and the green light, it just doesn't matter.
[00:04:35] Which brings us to this crazy paradox. If companies are reserving AI for their top performers and managers, they must think that's the best ROI, right? You'd think so. But the data says absolutely not. It shatters that assumption. Here's where it gets really interesting because, wait, they're not getting a return on that? Barely any. Let's look at the 2025 Bryn Yolfsen study. They looked at customer support agents using AI. Overall, productivity went up 14%.
[00:05:04] Okay. 14% is decent. Right. But it wasn't even. The novice, lower skilled workers saw a massive 34% boost. 34%? That's huge. That is. But the top performers, the highly experienced veterans, they saw almost zero effect. Wait, really? None? Almost none. And it's not just customer service. There was a 2023 Delacroix study out of Harvard Business School looking at Boston Consulting Group. Below average consultants improved by 43%.
[00:05:33] Wow. But the above average elite consultants only improved by 17%. Wait. So organizations are spending massive budgets buying AI licenses for their elite veterans who barely get a bump and actively withholding it from the rookies who could see a 40% improvement? That's terrible business. It really is. And what's fascinating here is the mechanism behind it. The researchers call it the novice advantage. The novice advantage. How does that work?
[00:06:02] Think about tacit knowledge. You know, that unwritten, instinctual stuff a veteran employee has. Like knowing exactly what tone of voice to use on an angry client? Exactly. Usually it takes a new hire years of failing on calls to absorb that, but an AI system ingests millions of transcripts from those elite workers, finds the patterns, and literally whispers that veteran wisdom into the rookie's ear on day one. Oh, wow. So it just captures the veteran's brain and gives it to the new guy.
[00:06:30] Right. It brings the bottom of the workforce up to the top. So the veterans don't get a boost because the AI is just telling them things they already know. Restricting access to managers is just a massive misallocation of resources. So if I'm an executive, I'm thinking, okay, I'll just buy licenses for all my novices and watch the profits roll in. But it's not that simple, is it? Not at all. Just throwing AI at novices is actually really dangerous. Because of the whole hallucination thing.
[00:06:57] Yes. And specifically what that Harvard study called the jagged technological frontier. The jagged frontier. Mm-hmm. What does that mean exactly? Well, AI is exceptionally good at some things, but shockingly incompetent at other tasks that look virtually identical. Okay. So it's like walking into a lake. You're wading in water that's perfectly waist deep for miles. You get super comfortable, right? You stop paying attention. And then boom, you step off an invisible 50-foot cliff.
[00:07:24] That is the perfect analogy. Because the water looks the same, you assume it's safe. When those consultants used AI for tasks outside its true capabilities, they were 19 percentage points less likely to get the right answer than people who didn't use AI at all. 19% worse. Mm-hmm. Just because they trusted the machine blindly. Exactly. They outsourced their judgment. But even if you avoid that cliff, there's another huge issue Gallup found. It's called the reward leak. Reward leak. Like, where does the save time actually go?
[00:07:53] Exactly. Say you use AI to cut a five-hour task down to two hours. What happens to those three hours? Gallup found that 31% of AI users are just being given more responsibilities. Oh, man. And for Gen Z workers, it's 38%. Their workload just goes up. It's like, instead of AI being a bicycle that gets you home from work earlier, companies are treating it like a treadmill that they just keep speeding up. Yeah, the time dividend trap.
[00:08:21] For everyone listening, seriously, think about that. When you save time with AI, did you get to rest? Or did you just get more tasks? Usually it's more tasks. That it turns this productivity miracle into a massive source of grievance. I can totally see that. So if the benefits are being hoarded by elites or just turned into more busy work, who is really getting marginalized the most here?
[00:08:45] The demographic data is pretty stark. There's a 2025 Danish study by Homelom and Vestergaard showing women were 16 percentage points less likely than men to use chat GPT for work. 16%. That's a big gap. And a global synthesis by Otis and colleagues showed this gender gap persists even when you equalize access. Wait, so even if everyone gets the software, the gap is still there. Why?
[00:09:09] It comes down to a lack of targeted support and perceived permission. Certain demographics feel less culturally authorized to take risks with new tech. And it breeds anxiety. Pew Research in 2024 showed 52% of workers are worried about AI. Only 36% are hopeful. Because they have no say in it. Right. 52% want more influence over technology adoption than they currently have.
[00:09:35] Okay. But here's my question. If women, older workers, and frontline staff aren't using these tools, aren't the AI systems themselves going to become super biased? Like, won't they just evolve to only serve the needs of male white-collar managers? Absolutely. Because models learn from their users. If only one demographic refines the tool, it only fits their work, leaving everyone else out in the cold.
[00:10:01] Wow. So how do we actually fix this gap? What are smart organizations doing? Well, the first step is targeted enablement. What does that look like? It means you can't just hand out a license. Walmart, for example, partnered with OpenAI to offer free certification explicitly to its frontline and office associates. You have to actively build the bridge. Okay, but what about the time dividend trap? How do they fix the treadmill problem?
[00:10:28] That's the ultimate fix. It's about redeployment. And we have two perfect examples here. The right way and the wrong way. Let's hear the right way first. The right way is IKEA. In 2021, they introduced a bot named Billy that ended up handling 47% of customer queries. Right. So that's a massive amount of call volume gone. Exactly. Now, they could have just fired those staff, but instead they retrained 8,500 contact center workers to become remote interior design advisors. Oh, that's incredible.
[00:10:56] Yeah. And by 2022, those upskilled workers generated 1.3 billion euros in new revenue. 1.3 billion? Just from reinvesting in their people? Exactly. Now, the wrong way is Klarna. The fintech company. What do they do? In 2024, they boasted that their AI was doing the work of 700 human agents. So they slashed their headcount. Chasing that immediate cost savings.
[00:11:20] Right. But by May 2025, the CEO had to admit the AI couldn't handle complex calls. Quality dropped significantly and they had to frantically start hiring humans again. Sure. IKEA's story is inspiring, but realistically, won't most executives just look at the Klarna playbook initially? Because a quick headcount reduction looks amazing on a quarterly spreadsheet.
[00:11:41] And this raises an important question because, yes, the temptation to cut is very high. And that's why worker bargaining power is the crucial final puzzle piece. You can't just trust executives to choose the IKEA route. Right. And before we get into this next part, just a quick note for you listening. We're sharing with the historical data and current events show about labor dynamics. We're not endorsing any political stances. Just looking at how workers are actually forcing these systems to balance out. So how are they doing it?
[00:12:10] They're securing massive contractual wins. Look at the 2023 Writers Guild of America strike. They ensured AI couldn't be credited as a writer. Which protects their jobs entirely.
[00:12:51] Wait, so the workers actually get to approve the investments? Yes. Yes. And in Germany, the Works Council Modernization Act explicitly requires an outside expert when employers bring in AI. So what does this all mean? It sounds like workers are basically demanding a technological prenup, right? Like before the AI moves in, we're going to agree in writing on who gets the house if things go south.
[00:13:20] I love that. A technological prenup. If we connect this to the bigger picture, whether you're in a union or not, having transparent agreements about consent, notice, and retraining is literally the only way to build trust. Without it, you get the Klarna disaster. Totally. So just to recap our journey today, AI has massive potential, but if we just leave it alone, it makes the rich richer and gives everyone else busy work. Right.
[00:13:45] We looked at how the novice advantage flips that script, giving huge boosts to rookies. We saw why giving workers a voice isn't just nice. It's mathematically better for the bottom line, like IKEA proved. And we saw how workers are demanding their piece of the dividend.
[00:14:01] Yeah. And I want to leave you with a final thought on this. If this AI dividend gap just keeps widening, what happens to our broader society? Are we actively creating a permanent, unaugmented underclass of workers who are just never going to catch up to the AI enhanced elite? And if so, what does that actually mean for social mobility in the 21st century?
[00:14:22] Man, that is a heavy question to end on, but so important. Thank you all for joining us on this deep dive. As always, evaluate how your own employer is rolling out AI. You have the right to demand a stake in the value your work creates. Don't settle for the treadmill. See you next time.
Sponsors
