Why Do We Rush to AI

Speaker A: Welcome back to Digital Value Creation. I'm, um, Arpad. I work for an AI focused hardware company.

Speaker B: And, um, I'm his brother Tomas, and I work for an AI focused software company. As many of you know, today we're going to do something different. We're going to explore something we've seen up close. The psychology behind AI adoption. Not the technology, but the human behavior that drives how people jump into the AI bandwagon or why they hold back.

Speaker A: Yeah, and it's funny, they always talk about AI as this technology wave, but when you look closer, it's really mirror human motivation. Actually, it reverses how we think about leadership, ambition, fear, and our careers. It's really not just about machine intelligence. It's really about human ambition and insecurity.

Speaker B: Ah, we have both. Exactly. AI adoption has become this new corporate Rorschach test, you know, the inkblot test in psychology. And some people see an opportunity, others see a, uh, threat. Like, we go, wow, this is exciting, or whoa, this is scary. Either way, everyone's reacting emotionally as much as rationally. So let's start with what's happening at the top. Because CEOs are the ones setting the AI tone. This is what we call the CEO AI pressure cooker. Arpad, how do you see this?

Speaker A: I mean, we see this pattern everywhere. Actually. Forbes recently published a study where they highlighted 73% of CEOs felt that their career might be limited if they really don't take advantage of AI, uh, if they somehow blow it. So CEOs really feel like they have to lead with AI, uh, not because they identified a business problem that only AI can solve, but because not doing it feels like you're just being left behind. And, uh, this is what some research refers to as the Gretzky Effect, where leaders are really skating where they think the puck is going, but the problem is the puck is moving way too fast, so they cannot really plan for that.

Speaker B: What happened to the puck?

Speaker A: Right?

Speaker B: And the irony is that many CEOs are implementing AI not out of confidence, but out of fear. They don't want to be seen irrelevant. And this is the social pressure at the top. If your competitors just had an earnings call and they talked about AI first transformation, your board will start asking, where is ours?

Speaker A: And then this is going to cascade. Suddenly there's an strategy office and there's a new chief AI officer, and the CEO start asking every town hall and mentioning that, hey, is everyone's job. And, and this is important. And it has been proven that CEOs need to drive this initiative. But without clarity, it is really just a blank check for confusion. So actually Gartner even predicts that 40% of AI project will be canceled and never meet production by 2027. So. But uh, ultimately this is not a technology failure. It's not that AI is not able to solve real problem. It's really a question of leadership misalignment. So Thomas, I know that's a key focus for you.

Speaker B: Yeah, I was at a meeting and somebody said, um, I think they said AI right now is activity masquerading as progress. People are busy, everybody's developing something, but nobody's moving forward. And it creates what I think Harvard called work slop economy. And I love that term. And you looked into this. So what is workslope?

Speaker A: I don't know whether it's HBR who claimed it, or somewhere else, it already showed up. But they coined this term workslop. And I find it fascinating because what they find that 40% of the employees today receive AI generated presentations. Content that looks polished but add no real value. And that's what they call workslop. This is the AI slop that we saw on the Internet, but in the

Speaker B: workplace, uh, it's the new corporate version of AI spam. And I can tell you I have a full folder full of AI summaries of meetings that I never open. I have to do lists by the hundreds after every day that people are sending me of all the meeting action items, slides that people are generating. They look ugly, they're not even on brand or people draft an email. And I can tell that I had one question and the answer is four paragraphs long with the double hashtags. The content is off, it's generic, it's missing context. It actually creates more work for me to understand whatever they're talking about. And then trying to fix it doesn't work right.

Speaker A: And that's really shift the burden downstream. So all of a sudden the sender looks productive, I mean, and the receiver has to clean up this mess. And this is actually what the research quantified, that every instance of working and this workday slop cost nearly M2 hours of wasted time. So there's a estimated 9 million dollar per year in a 10,000, uh, strong company. I mean if you think about how many, as you said, how many meeting summaries now we can create that nobody really reads there. We are not really extracting what actions or what decisions needed. You're just sending it out without a whole lot of uh, thought process behind it.

Speaker B: So that's the negative ROI of AI so far. It's like crazy, uh, and it's incredible. But this is the human part. None of us are doing this intentionally. People are doing this because AI work has become a career signaling mechanism. If I'm doing this, I'm with AI. I'm doing stuff, right?

Speaker A: Exactly. And if your company is pushing AI, you want to be seen as a part of the story. So people churn out AI generated content not because it's useful, but because, hey, it's politically smart. I want to make sure that I ensure my career, that I use the technology.

Speaker B: So, uh, somebody called it hashtag activism. It's like, look, I'm innovating using AI. Hashtag using AI. And yet the study you mentioned found that 53% of the people feel annoyed or even offended when they receive a clearly AI generated junk work. So we are eroding trust in the company, we're eroding trust in each other, which is really problematic. And that takes us to this big topic you and I often talk about. Is the hammer looking for the nail?

Speaker A: What is that for the backward problem solving. I mean, this is really not new to AI. Uh, I think I just accelerated. I mean, we all seen this mistake when a new technology arise. All of a sudden we are excited about what problems can be solved with this new technology, how this tool can help. And people start with, hey, we need an AI, uh, use case because I have this capability. Or, hey, a new model just became available. It's multimodal. What can I do right now? Where can I stick it on?

Speaker B: And this hammer thing, like you said, it's with every technology. This is not new. It's happening at AI scale. When you turn on the automation that's built into AI now, it's coming from every angle because the barrier of entry, of turning on AI tools now is almost nothing. So it happens everywhere. So I know you mentioned an interview you listened to.

Speaker A: Yeah, it was fascinating. Uh, Connor, who is a chief architect for New York University, and he really entered this field and become a thought leader from his angle on psychology, not technology. And he really focuses on how adoption even more challenging than the traditional technology challenges we already seen. Because as a generic technology, whether it's Jenny I originally, it can solve many things. It's general purpose technology. So all of a sudden it creates this use case treasure hunt. And, uh, we just see, uh, mushrooming use cases to see, look what is possible right now, what can be solved. Yeah.

Speaker B: So I think in the prior episode we, we talked about the MIT study, I think by now everybody talked about the MIT study had 95% of, of AI, uh, projects fail. And it's missing many points. I think one of the points is we'll come back to why it fails. Partially because it happens in a vacuum, in a hybrid tower, in a lab. But the other reason is because you're not actually solving a real problem. In fact, you're not even looking for a solution for a problem. You're looking for, like you said, um, um, a problem for a solution. And it's always backwards. So it's not because AI doesn't work, it's because again, we're not asking the

Speaker A: right questions or we really settle for a lot of small use cases that solve somebody's problem. And hey, that's great, but are we really focusing on what makes a difference? I mean this is where, and I think we mentioned that last uh, episode as well, uh, when MacKenzie focused on organizations that were on the 5%. The 5% are really had breakthrough results with AI, but they find that they actually took the time and completely re engineered complete workflows. They looked at the process end to end. And this is where you have seen significant value creation because all of a sudden you are not just chasing the next use case, you look back what were ineffective in your organization. So you can really double down.

Speaker B: Yeah. And it's hard because it means breaking silos, rethinking initiatives and frankly challenging middle management structures that existed for decades. So if you find these use cases that are truly new, that are solving unsolved problems, then you're going to start breaking things. And that's hard. So frankly, it's easier to start a, uh, flashy aicoe than rewire your exec in the entire business process. Right. So many center AI centers of excellence, reported the CEO, which sounds super powerful, but that isolates them for real operations. They become these ivory towers or superlabs of experimentation with no muscle memory of past projects or not even real experience of execution in the field.

Speaker A: Right, right. And this is when you see the vanity metrics focus on usage and adoption instead of value creation. And it's almost like the R and D departments that never ship any product. So instead of scaling a solution, companies end up scaling presentations about AI. And that leads us.

Speaker B: So this creates this. Yeah, so this is uh, I wanted to jump in the expert paradox. Right. So you know, like, it's really interesting because we talk about like, you know, there's these labs that don't really have the expertise, but then we're questioning do we even need the expertise now with the age of AI?

Speaker A: Right, Yeah.

Speaker B: I think that's philosophical.

Speaker A: It is, it is philosophical. Because if I can teach you any skills on demand, whether it's coding, analysis, presentation, design, is expertise still relevant? Um, is it, is it still important?

Speaker B: You know, we read so many studies, uh, we'll, we'll put it in the um, links. But one of the study I remember talked about AI tutors. And the AI tutors, it's a really interesting challenge. One thing, you can learn anything from using AI, right? You can learn the basics, the text, the subject matter, expertise. But typically what happens in business is somebody's holding your hand, somebody's helping you with your workflow. Now there are studies on AI tutors and they turn to be as effective as, as somebody who's your buddy, who's your work buddy, who's your manager, who can guide you or your trainer. In one experiment, people learned problem framing, which is understanding the problem and figuring out how to solve it, 32 fast, 32% faster with AI tutoring than with classroom training. So this is fundamentally different, right?

Speaker A: It is. And um, I think it was, it was another study built on that that highlighted that many companies who first rolled out, I noticed that some of the mid level, um, experts all of a sudden rise to the top because they might have been faster to adopt AI and uh, an almost outperform, at least in a short term, they're seasoned, expert. But when they followed up, they realized that that early shine actually started to dim because there is a paradox. You can upskill, uh, anybody, but there is a differentiation between knowledge and judgment. And AI can teach you what to do, but it might not give you the instinct about when to do it and why it matters. So human experience still is critical. And that was a key takeaway from the study that don't dismiss your experts, uh, who have this intuitive judgment because, well, you can have quick knowledge demonstrated, but these instinctual skills are still very relevant.

Speaker B: Yeah, exactly. Although this is one thing we don't know. Right. So all of this will change the meaning of mentorship tutoring. Uh, so we may not need a senior person to show us how the system works. So we may need a model on how to think. And can that model be taught by an AI? And if it can be, that's a fundamentally profound cultural shift in an organization because part of the reason we have a hierarchy is for this enablement and tutoring and teaching. So is that gone?

Speaker A: Right, but that's what raises an important question because, uh, and that frankly creates a lot of anxiety right now among experts because expertise has these two sides. One is knowing something, but second is being able to call judgment calls. And if AI can really replicate domain knowledge, then really expertise start to feel like a depreciating asset. And that fear might actually lead to resistance to adoption. Especially in many, many senior roles.

Speaker B: Yeah, I mean it's not irrational. Right. So if AI can code, it can draft, it can summarize, it can design workflows, it can tutor my employees, the question becomes what's my unique contribution as a senior person? So we've moved from man versus machine to manager versus meaning, whatever that means. Right. And that fundamentally changes entry level role. So we talk about the senior leadership, we went from CEO to senior leaders. But it changes how entry level uh, people enter the workforce. So what's about that?

Speaker A: I mean we both have uh, daughters and sons and I think we all see that already in the marketplace because there is a question, quiet crisis that is happening right now. I mean some studies says up to 26% of uh, of some of these high paying entry level jobs, especially in technology and marketing, are either disappearing or getting a significantly more selective. Because this entry level research shows that um, automation and AI really eliminating a lot of the groundwork are also. There's a lot more to that than just getting the work done. That what people used to learn the craft, to understand how to get things done, to develop that instinctive. And this is going to disappear as well.

Speaker B: Yeah. It's crazy, right? So my daughter just entered the workforce. Um, and consulting used to be the place where you go as an entry level person and they would hire thousands and thousands of people. And that almost completely dried up. The big four is not hiring. They're not hiring and they're not entering this across. This could be. Law firms could be a lot of different roles where they used to be the source of new talent. So they're saying, hey, the kind of work we used to give them now can be done with AI. So I think um, in the olden days, the first few years for somebody was building judgment through repetition, cleaning up data, creating drafts, PowerPoints, formatting slides. Now if AI does all that, what can entry level people learn? What kind of tutoring can we open up for them? Right. So how do you think about that?

Speaker A: They both had a consulting career before we shifted to industry. And uh, that was the ultimate apprenticeship model. And that's what the research points out that that is that we start to lose and organizations risk creating that judgment gap where a Lot of senior expert can be replaced eventually with a lot of automation for that knowledge. But if you are not developing entry level workforce to learn by doing and learn by making mistakes and being mentored, um, we create this development gap. Yeah.

Speaker B: Ah, it's going to be interesting uh, because there's a point like maybe there's not as much use for senior leaders, then maybe there's not much use for entry level. But now we're finding out maybe there's not as much use for mid career people that have this, gained this experience, they can, they can guide the rest. I have no idea what we're going to do. It's like the whole organization is being like every role is being questioned in terms of, and they have to rethink themselves because we're not seeing massive shifts in the workforce yet. We're just seeing all these questions pop up.

Speaker A: Right. Uh, and maybe I'm thinking about this as the same as we are thinking about how we train models that you can over optimize and sometimes over train models and uh, that might lead to very selective problem solving. So it's very easy to jump in a bandwagon that hey, let me just cancel all entry level jobs or let me eliminate all the knowledge work. I think organizations need to step back and say how can I maintain some level of diversity, how can I maybe reshift, uh, what these roles need to do and fundamentally stay adaptive. And I think that leads to a lot of hype like this really balance between jumping to the hype and hesitation to stay at the sideline.

Speaker B: When we started this episode, I was all hyped and now I'm all hesitation like we're talking ourselves into skepticism. Uh, but those are the extremes. Either you're an AI evangelist right now or an AI skeptic. And both groups are motivated by emotion, vision. Most of us don't have actually proof and data. And um, I think the evangelists are driven by status and curiosity while their skeptics are driven by fear. Fear and maybe fatigue, maybe the work slot fatigue.

Speaker A: Right, right. And um, maybe both are right. I mean if you look at AI enthusiasts and even though today's episode we talk a lot about the risk, we both believe this is a transformational technology. And we also help our teams because we see that this is their chance to refine their career. We actually help our daughters and sons like this is the chance to redefine a career because we see the possibility, but we also see the cautious ones asking that hey, what's the catch? Are we creating just Another bubble like um, metaverse or web3 or um. This is a question of how can we really focus on what matters.

Speaker B: It's funny, I was just talking to um, a group of friends about um, Metaverse because people are saying meta now is all in, in AI, of course, massive investment from Zuckerberg. I remember when it was the same thing about Metaverse. It was going to be the big bet Facebook was made. There's nothing wrong. I mean this is just the nature of technology. Massive bets from technology companies and massive bets from all of us. Like we, we lean into these things. We're leaning in with our skills. So there's truth to the enthusiasts and probably truth to the um, uh, warnings, um, the skeptics because we've seen this song and dance play out before in boxes NFTs metaverse so we don't quite know. This time is different. But you and I remember we heard this so many times. How it's going to play out we don't know. Um, so we're ultimately, I think the consensus is, and you and I believe in it, we're entering a fundamentally different shift, although we can prove it, that's the new industrial revolution. 5th, 6th now I can't remember. But history also tells us that early movers often struggle. In fact, if you look at who won the dot com battles were not the early movers. It wasn't Yahoo, it wasn't altavista, it was Google. Right. So this may happen here. So some of the early movers may not be the finalists and, and ultimately the ones that win make real, create real value once that, once the hype

Speaker A: clears and I think the best leaders we talk to, that's what we are there. And they are not blindly pro I, they are strategically curious. They experience and look at how can they scale and how can they focus on areas that worth investing. And they understand AI isn't just a tool, it's almost like a mirror. It reverse what your organization doing great with clarity where they are lacking. Yeah.

Speaker B: So when we look at AI adoption we talk about in this episode about psychology. So on one end the opportunists will um, look at AI as a letter and the perfectionists will look at, they'll wait until it's built, until it's ready, until this is the roi. And the winner will be somewhere in the middle. People who understand that it's imperfect but imperfect actions will beat perfect execution. Right. So that's where psychology plays a role. The middle will ultimately uh, define where A.I. goes. Um, so let's talk about something different. How is this redefining our leadership? How do you think about that?

Speaker A: I think we touched on it throughout this, uh, during this point that AI fundamentally challenges what leadership will be. Several research studies argue that success now depends less on technical know how, less on your knowledge, more about how as a leader, you can span across boundaries and redefine organizations. Back to the McKinsey point that as an expert in a domain, AI can actually help people to get there. But as a leader who can connect the dots across parts of the organization, maybe parts of your whole ecosystem, that's what really make leaders rise.

Speaker B: What I love about this AI movement now there are new words never heard before. So one of the words we heard is boundary spanners. How leaders will break down organizational boundaries because AI doesn't fit. AI is not a departmental technology like erp. You had a finance system lived inside finance and a CRM lived inside sales. AI works across boundaries. So it becomes sort of a new language. How organization works across. So it's not a department, it's a new language that everybody in a company may speak. And, uh, ultimately, how does it change the CEO and everything else?

Speaker A: Arpad? I think it definitely changes all the C level roles because instead of asking what's our AI strategy from a CEO's seat or for a department level, they really should ask, who do we need to become, what kind of organization we need to be to use AI effectively? Uh, and this is really a mindset shift from owning technology to really orchestrating this. As you said, boundary spanner, transformation.

Speaker B: Boundary spanner. I love it. So let's bring this plane back home. Psychology of AI, huh? Um, if we had time to write a book. Oh wait, AI could write a book. Anyway, so if we strip away, stripping away the jargon and the hype, what we're really seeing is a new social contract. I read it somewhere. New social contract between people and technology. So AI will not replace humans. Let's hope it will amplify the best and worst parts of human psychology.

Speaker A: Right?

Speaker B: That's what we're hoping for.

Speaker A: Absolutely. I mean, the future of digital value creation isn't about replacing human intelligence. It's about redirecting it. I mean, AI gives us leverage, and we can use this leverage without purpose, just to create action and activity and just scale mediocrity faster. Or maybe we can redirect it.

Speaker B: So we talked about, uh, are we solving the right problems? So maybe one of my takeaways from today is I think we could all ask ourselves, our employees, our peers, our colleagues, are we using AI to make work more meaningful or just to make ourselves more visible? So let's hope for the former. Right.

Speaker A: Perfectly said. Uh, I mean hopefully we don't create more AI slope or work slope and really that's our challenge and that's a challenge for us and for everybody listening today. Don't just use AI. Understand what it reveals about you, your culture, your values, your team's values and how we will link um, this to measurable outcome. We will share. We quoted a lot of research and a lot of report. We will put that in links because uh, uh, this is an ongoing discussion and today we wanted to move away from technology and pose these questions. Some of them the answers are still being defined and would love to hear from you. What do you think?

Speaker B: Absolutely. So thanks for joining in today. Thanks for listening to Digital value creation. Until next time, stay curious, critical and most importantly stay human.

Speaker A: Thank you.