Lenovo’s Global CIO:
Advice on Scaling AI, Managing Global Uncertainty, and AI-nomics

Lenovo's AI Strategy & Impact

Michael Krigsman: How does one of the world's largest companies think about scaling AI, making AI investments, and the impact of AI on chief information officers? Today, on CXOTalk number 887, we discuss these questions with Art Hu, Global CIO at Lenovo, who is also the chief delivery and technology officer of Lenovo Services and Solutions Group.

Art Hu: The chief technology and chief delivery officer role for our Solution and Services Business Group is the newer one, in addition to my global CIO role. I am privileged to be at the leading edge of Lenovo's services and AI transformations.

Michael Krigsman: Lenovo is one of the largest technology providers in the world. Where do you stand with AI today? It's a very general question, so your perspectives.

Art Hu: The upcoming decade will be very much around hybrid AI, and in all its incarnations. So if we think about, as a technologist and a business practitioner of technology, meaning that I really focus on, of course, exploring the leading edge of what's possible in technology, but helping land it in a relevant way for the business, I really try to chunk it up and look at it in domains.

The first one, for example, around innovation and product development. We're selling more than four devices a second. We have AI services and solutions. And so when we think about the product development, we're really bringing that into all aspects of our portfolio. So the AI is coming to our PC experience, our edge experience, our phone experience, our server experience, our services and solutions. So that's around product development. AI is coming into our entire portfolio and offering in way that's gonna be more and more integrated.

It's also coming into our innovation process. We're accelerating the discovery and product development on both the hardware and software and solutions side.

And then the last area that I think about is around our operations. So, and that, again, is all aspects of our end-to-end value chain. All of our functions, all of our geographies, legal, marketing, sales. I mean, for example, we actually have an enterprise seller AI coach, who can improve your selling motion. But HR and our supply chain is one of the top supply chains in the world according to Gartner.

So like you said, I wanted to do an overview. Whether it's innovation or offering portfolio or how we run a world-class operation, we're finding ways where AI is relevant. And that's what's interesting. It's relevant everywhere. There's nowhere you can point in the company where someone says, "You know what? That doesn't apply to me."

AI Investment Strategy and Agility

Michael Krigsman: How do you think about AI-related investment decisions?

Art Hu: It is advancing at such a rapid pace. Whether it's... And you can count this almost by day or by week or by month, but the pace is incredible. With that being said, I still think it makes sense to have a framework, because what you don't want to do is suddenly feel like you are lost or that you're swimming and drowning and too much that's happening.

The best way to avoid this analysis paralysis, because it's very easy. Huh? What's best? Should I just keep waiting because the frontier seems to be moving so quickly? I'm very much, and at Lenovo, we're very much about learning by doing and tolerating ambiguity.

So when it comes to our AI investment strategy, it is very much where it is a full-court press. Everyone in the company, and this is not just me and any of my roles, but it's the entire executive committee, we have a joint commitment to make AI a key part of our business, because we fundamentally believe that it's going to bring the return.

And so in terms of the investment question and our investment strategy, the overall is everyone is expected to engage, and that helps tremendously. There's no need to explain what we're doing or why, and that really helps the team step forward.

And then the other part of it is that we help, from a technology perspective, we help inform the business on when should we open the aperture on exploration. We'll have many parallel experiments. We're gonna see and test what is the boundary, and we're gonna have times where we come back together and want to do a little bit of summarization and harvesting of what we've learned so that we can focus and really apply.

So I think the couple of key parts is we're all in and we're all in as a team, not just as a technology team, but with the business teams. And then we have to be flexible. Yes, we use what's worked in terms of having some framework to evaluate, but recognize that it's not gonna work perfectly and there's gonna be some ambiguity.

For example, in areas that are more mature, we might not want so much ambiguity and parallel processing and maybe having overlap, because why overlap if you already know what you're doing? But in a field that's changing so quickly, we explicitly have to say, "It's okay if we do things that seem duplicative," because we might find different things or different approaches towards similar, seemingly similar things might generate different results when the teams come back together.

Michael Krigsman: As you mentioned, AI is changing so rapidly. The technology is evolving, and the usage inside organizations of every size. And so how does that factor into your investments? I mean, how do you ensure that you're picking the right horse when it comes to specific AI technologies, for example?

Art Hu: We will want to find investments that are no regret, meaning that we are going to get value from them with the business in the window we're wanting. And that makes it easy, because if we've gotten a value on a fixed window, then even if that doesn't end up being the long-term solution, then we're still getting a return on investment, which is responsible for our shareholders and our stakeholders. And so in that respect, that's from a financial management perspective. You can still find investments that everyone's excited about because it generates that return.

At the same time, there's no guarantee, and I think that's... can be very scary for a controlling function or if you want that absolute certainty, you will definitely not find that here to say the AI solution we picked today, whether it's the model or it's a gateway or whatever it is, it's a quantization method, none of those things are guaranteed to be cutting edge in six months or 12 months. That just doesn't exist.

But I think the point is, around no regret, what we have to think is you have to keep moving forward. And so the fundamental response, because another way of saying things are changing very quickly is uncertainty. That means there's uncertainty, there's no guarantees. But what we believe at Lenovo and what we're practicing is that the best strategic response is agility.

And that, AI definitely affords, the way that it can help plan better, the way that it can help get more information, the way it can help get more varied data, the way it can help simplify processes, all those things come together, whether in the micro or at the company or the department level, are things that you can use to create agility.

And I think that's what's so important, which is if you can reframe some of the strategic discussion, where investing in AI is actually going to help your agility and that's of strategic value for the enterprise, that's a very worthwhile discussion. Because that clears the deck about, "Well, how do we know?" 'Cause it's very natural to say, and we've seen, things that were cutting edge a year ago are now no longer cutting edge. They're commoditized.

But if you've gotten the value, if you've already helped increase your AI index for the company, that's okay, and then you can continue to explore. And the point is to not lose momentum and to continue in the face of that uncertainty. That's what's so important.

Michael Krigsman: We know that AI is fundamentally important, requires investment, and organization of the team behind it, as you were describing, even if we don't know the exact shape of the technology. And so therefore, agility enables you to invest strategically but essentially keep your technology options open. Is that an accurate way to describe what you just said?

Art Hu: Yes. And then the final point I would add on top of that, on why the mindset and having the right mindset for this is so important is, this is one of those things where you may not know what the next mountain you need to climb is. You may not know what the road looks like beyond the next bend until you actually get there. You can guess, you can simulate, you can approximate, you can estimate, but the point is, unless you move forward, you won't know. And only by moving forward can you get to the next.

There is definitely not the, "Please provide a three-year roadmap and a guaranteed vendor list and techniques and a budget that's going to fit down to the dollar that we can reconcile."

Michael Krigsman: Folks, right now you can ask your questions of Art Hu. If you're watching on LinkedIn, pop your questions into the chat, and I see some questions are showing up there now. If you're watching on Twitter, on X, then use the hashtag #cxotalk. But use this opportunity, take advantage of it. When else will you be able to ask the global CIO of Lenovo pretty much whatever you want? So take advantage of it, ask your questions.

Technological Evolution and New Possibilities

Michael Krigsman: Art, in addition to the technology shifting, right now we have tremendous economic trade, tariff changes, and shifts. How does all of that play into what you're doing with AI, how you're thinking about it, and how you're making investments, and also your customers as well?

Art Hu: On the technology side, I think it is really a seminal moment because there are now new possibilities. So at Lenovo we have the world's best hardware portfolio, and now we're enriching that with services.

But even for us, the new possibilities, because historically we've had, in the last few decades if you think about it, we've had a very fixed paradigm. From the dawn of the modern computer age we had decades where the command prompt was the primary way of interacting. Then we had the graphical user interface, the GUI, and that's been going for probably four or five decades.

But now we really have, for the first time, a way that some of the augmented reality and voice, how it may change the interaction with compute. We went from command prompts to GUIs. Voice has always been hanging around there as kind of on the peripheral, but we never got there quite because of the inability to really understand intent and have multi-turn and true human dialogue.

But as that becomes more and more capable with the latest generation of technology, the way that we interact with compute will change. And so that evolution is tremendously exciting. This has not happened for decades, where there was truly something that we could be on the cusp of.

And that's so interesting on the technology side as we transition into Michael, what you were asking about the economic, is because this starts to make things that were not really possible to do, or really just impossible to do, into the possible.

Now, to be clear, I don't wanna be hyperbolic. We're not talking about physically impossible things that are now possible, but technology capabilities in a business and a consumer context are opening up new frontiers.

If we think about the real-time capabilities and the ability. We see, for example just as a very simple example, if we go have one of the leading language models go into deep research mode and say, "Hey, I would like a master's level thesis that is 50 pages," it can go off and work on that for 30 minutes, 60 minutes, 90 minutes and come back with something that, you know, you can say 50% there, 70% there, but something.

And that was not possible five years ago. And those are the types of things where, as they become possible, they open up all these new frontiers, whether you're in a consumer space or a more business to business and enterprise space.

Michael Krigsman: Does that make sense? It does.

The Importance of Reskilling in the AI Era

Michael Krigsman: So let's jump to some questions. I love taking questions from the audience, this audience is so smart. And let's start from LinkedIn, and there's a question from Ravi Karkara. And Ravi says, "In an era where AI is reshaping every function across industries, why is it mission critical for corporations to invest in skilling, upskilling and reskilling their workforce at entry, mid and senior levels, not just to stay competitive, but also to drive responsible and strategic adoption of AI across core business functions?" So, the importance of investing in skilling and reskilling, but at every level.

Art Hu: The importance is simply that we still run the businesses. And with all this change, and as part of the business, if we are not changing then we're being left behind. And so I think the imperative... I, at this point I run into basically nobody, everyone would support the imperative that reskilling is necessary.

But I think and in some sense, Ravi, if I could try to get underneath the question. So no one would disagree. If you go to any chief HR officer and you're like, "Oh, we need to reskill for generative AI and the Agentic and the hybrid AI era," I think 99 to 100% people will agree. But I think the more interesting question, is how do you reskill? Because there's different ways to go about it. The fact that you do need to reskill for entry level, mid-level, as well as senior and even the C-suite is absolutely there.

But you need to take different paths. So let me give a couple of examples or try to talk in more specific. I think from an overall approach perspective, what does not work is a top-down. The office of the chairman and the office of the CEO say, "You must learn AI now. Here are the 10 courses, and you must be certified by the end of the quarter or you will not get a good bonus," or something like that. You can imagine something that's very top-down and driven.

I think this ties very well with what Michael teed up so wonderfully in the questions, which is this notion of uncertainty and the pace of change. That has to inform the approach. And so the approach has to be much more organic where the company is not mandating, but the company is encouraging. The company is creating the fertile ground and the environment that invites employees in to explore and to change.

Because the reskilling that happens when you are exploring because you think it will make you and your team deliver better outcomes is way better than someone sitting in a central COE trying to guess at what you need. And so I think that principle of the best thing the company can do is to create the preconditions and the right environment to go reskill.

So for example, at Lenovo, we have built hundreds of agents, we have hundreds of environments, all approved in case my security team listens to this at some point. But we are creating the environment that invites people in. If you want to work in legal, if you're in marketing, if you're in finance, if you're in HR, there's something for you to work with that will help invite you in the door. And what we're finding then is there's a pull. People will say, "Oh, I have already learned these things. Now I would like to learn more." And that's a much better way in terms of what we've found.

So we approach it not as a, "You shall or shall not do these things," but creating the pull, creating the impetus so that the employees are leading the charge, and they pull the company behind them. And I think that approach works equally well whether you are a firstline employee, a firstline manager, a middle manager or an executive.

The way you do it, of course, is different. But the whole point is you have to spend enough time getting hands-on with the technology, and this is then as you start to become a middle manager or as an executive. The final piece I would say is it does take an additional point to read, to go use, to try out some of these tools to see how powerful they are because it is intrinsically my belief that it is impossible to go work on and to effectively lead the technology if you don't know what it is.

And so the more senior you are, because it's not as natural, you are not going... I, as a CIO, I'm not coding 100% of the time, and any CIO who is probably... Then why are they the CIO? But the more senior you are, the more it's important that you spend the time to understand what it is by getting hands-on with it, by understanding what is at the frontier so that you can lead and create the right environment for your teams.

Michael Krigsman: Go to cxotalk.com and subscribe to our newsletter. And we want you to come back, audience, to our upcoming shows. We really do have tremendous shows.

Overcoming Fear of AI Replacing Jobs

Michael Krigsman: How do you overcome the fear that many employees have that, "Hey, AI is going to take my job and I'm not sure what to do about it?"

Art Hu: The framing is what is less than constructive, we'll say. All right? Because so I've got two issues with it. One is just from a personal philosophy, it's very passive. We talk about agents, well, let's go back to what the root of agents are. It is agency, and agency is a fundamentally human thing. I am... And yes, there's, you know, constraints, but fundamentally I am able to take decisions and actions that will change my path. That's what agency fundamentally is.

So just from that framing, "AI will take my job," makes it sound like you are sitting there, a static entity waiting for something to happen to you. So I already tend to reject that framing or in its most simplistic form.

If we then think about, well, what else? How can we reframe it so it's more constructive? I think it's more, "AI is coming, so how do I understand?" And how do we step back and make it a more constructive framing, which is, it's a technology and it's going to make its way into everything. You can dispute the ticky-tack stuff is at three years, five years, 20 years. I think we'd all be shocked if things in 15, 20 years looked at all the same that they are today.

But the final part of this is that AI doesn't take jobs. What it does... It's a technology that automates tasks. Jobs are still fundamentally designed by people. Because it's a very human enterprise of what is this company trying to do. And so you understand that AI is just automating or augmenting some tasks.

By embracing that, you are taking agency back onto yourself because then, it's not about, "I'm waiting for someone to take my job. No, no, now what do I do?" It's much more if you are in the curiosity and if you've been experimenting that it's much more about redefining the tasks that compose your job. So yes, you can... In the narrow form, AI is taking tasks and making it better, but those are probably not the tasks that you wanted to do. Did you really want to go to the website, download the report, put it in an Excel at the sa... Like, you don't. You can automate that, and you can actually do that much more flexibly.

But if you don't have to do so much of the grunt work, and by applying just this very classical approach to problem solving of decomposition, it's much more manageable. All right.

The Evolving Role of Software Engineers in the Age of AI

Art Hu: For... Let me just... Again, let me make it very specific with an example. If you are... So developers, if you are software engineering, it's very easy to even get unsettled because, oh, all the headlines you see are about, in the future, this company doesn't want to hire software developers. Oh, you can just write code. Oh, how many more lines of code can you write? My code is 100% faster with people, and now, it's already as good as a junior developer. Next year it'll be good as a senior. So, it's easy to get swept up in that.

But if you go back to tasks, we all know... And this is one of my pet peeves, which I'm always trying to help educate more, many senior leaders believe engineering is about just writing more code. Kind of the Hollywood trope of someone who's at their computer, room is kind of dark and your fingers are flying over the keyboard and maybe smoke is coming off of them.

But as a developer, that is maybe 10 to 15% of your time. There is thinking about design. There's testing. There's regression. There's performance. There's architecture. There's security. And so just let's take even one part of that, the whole design.

You can now as a developer do things that previously you might have needed a user experience designer to do. If you wanted to create a prototype, previously you might have worked with a prototype specialist and worked on something online in one of the modeling tools. But that would have taken time. Now you can just say a few words and a few minutes later you get something that is a starting point. Maybe you can even show it to the business.

And so that then helps you reframe. Something that you couldn't do before, you can now get almost for free and almost in real time creating a workable demo, and that helps you go back to think about your job. Are you really just there to write code and get requirements? No. The fundamental point of software engineering is to deliver an outcome for the business. So as long as you have the framing right, all you've done is you can do more, and you have better tools to do so, and now you can redefine your job to focus on either broader scope or more value-added components.

And so I think the point is... And again, I think history for everyone who is concerned about that, it is not without reason because in the moment it will feel like someone is coming for your job. But for all of these technologies that have come historically, it is always the things that are created afterwards.

Certain jobs, if you define it very statically of it can... Like a software engineer just needs to write code and run the test case, then yeah, I would agree, that job is not gonna really exist in that exact form in three years, five years, 10 years.

But having a valuable software engineer who really understands what the enterprise is about and can fully use the tools to create something. Maybe it's not even a classical application anymore. Maybe it's a chatbot. But the... If you're thinking from that, then you're still a software engineer. You swapped out 3, 5, 7, 10 of the tasks that were more laborious. But you can define for yourself what that is.

So... And I've had this discussion so many times with my teams because software developers, if you read the headlines, you'd think they were an endangered species. But this is fundamentally something that's much more constructive and much more empowering for the AI is coming for my job.

Michael Krigsman: This is an excellent time to subscribe to the CXOTalk newsletter so we can notify you of shows that we have just like this. We have amazing shows coming up. So just go to cxotalk.com and subscribe to our newsletter and do that right now.

AI's Impact on Industries and Regulatory Challenges

Michael Krigsman: We have a question from Preethi Narayan on LinkedIn, and she says, "Which sectors are leading in AI innovation and what's driving their success?" It's an interesting question.

Art Hu: There's a quote about economists saying that they see the impact of technology everywhere but the numbers. And so what I mean by that is I think given the very early stage, I think it is in general going to take some time before you can see true acceleration, and relative acceleration in macro, economic data, at that level, or something you get from the National Bureau of Statistics and other official agencies.

That being said, I think, in general, industries that are highly regulated, will have to move a bit more cautiously. And that's just for the very simple reason that, the laws, the regulations, the compliance requirements, will impose some additional process overhead and boxes to go check. And, you know, things that are around sensitive data, critical infrastructure, personally identifiable, private data, any scenarios that touch those will necessarily need to be a bit more... Well, they'll need to invest more again in the compliance. So I think that's it as a general rule.

I also, not being a macroeconomist, I would also love to see, the statistics as they come out to see what might be some leading indicators, where AI is being applied. What we do see, though, is very practical use cases, and I think many of them are familiar. But fundamentally, these are the ones that are leaning into the strengths of generative AI, around creativity, around the ability to understand intent better than any class of technology previously. So that's one aspect.

Understanding Generative AI's Strengths and Weaknesses

Art Hu: And I think the other one is not only using the strengths of that, but also being tolerant of its shortcomings. And I think that's particularly important. If we shift from kind of the macro and sectors, because there's some general conditions there that I talked about. But then if we look at which use cases are going to make more and where are they going to cluster, it's not only it uses the strengths, but tolerant of the weaknesses.

And that's very important. If you think about, because then the weaknesses being, of course, is that the current architecture and class of technology that powers generative AI around transformers, will hallucinate. And you can say, is that a feature or a bug, because it's meant to approximate how human brains work, with some of the technologies and propagation and weighting of neurons and cells. So, you can argue if it's a feature or bug, but the reality is that kind of happens.

And so if you look at the areas, for example, around marketing, around things that, around human talk scenarios, where you can still do something very meaningful but have a human in the loop around contact centers, all those things really tap the creativity. So around marketing, around writing press releases, those are things where getting the wrong comma, having one wrong word, it's highly tolerant for that, because it's already 100X better and 100X faster than what it was before. And so people will say, "Okay, well, you know, if it's only 90% as good from an end product, but it's 100X faster and ten times cheaper, that's a win-win, in many people's books."

And so I think those are the areas, 'cause we are also continuing to search. We are still very much at the dawn of who can go more quickly. But I think ultimately, 'cause I think the question is very important, which is around speed. I think everyone wants to use generative AI for their enterprise to find that additional velocity that they can get for their business.

And so a search pattern I could offer is, is that, how do you continue to lean into and continue to identify cases in your company, in your context, that use the good parts of generative AI and are tolerant? Not that, you know, you never want to accept, but that are more tolerant of some of the drawbacks, primarily around the fancy word is stochastic, but the human word is random, randomness that is innate to the technology.

And that, and I think that's ultimately and we can talk a little bit more about it here, Michael, I'll give a stub, but essentially, that's going back to the you have to keep moving, because only by continuing to move can you compound the advantage. It's very unlikely a single project will take you from a middling performer to a world-beater company. It's gonna be the compounding of finding these use cases and continuing to move forward. And so it is super important to have some search criteria in mind to help sift through and identify what those use cases are for you.

The Role of Technology in Business Contexts

Michael Krigsman: It seems to me you're always bringing the technology back to the context of what the business actually requires. So it's not just technology in abstract. It's not just change in abstract, but it's what's going on, what's the impact on the business, and how do we have to adapt as the business and the technology and the environment are all simultaneously changing and evolving.

Art Hu: You've just summarized a lesson, it's a bit of a bitter lesson in the sense that as a technology industry, we learn repeatedly, because even until today, as you said, if you do just technology, shorn and stripped of its context, or you say, "I'm just doing the technology," without a clear view, this is how you end up with large project failures. And even today, you see in the news, you can just search, "ERP failures," and you will see from the public sector to the private to government, we still see projects that are exactly making that mistake that you just summarized, Michael.

Michael Krigsman: We have a couple of questions now, which I'm gonna combine, from LinkedIn and Twitter.

AI's Impact on Leadership and Adaptability

Michael Krigsman: Greg Walters is asking about the impact of AI on the C-suite. Specifically, he says he thinks that ultimately, AI removes the need for the C-suite. And on Twitter, Arsalan Khan is asking about the role of IT when it comes to AI, AI discussions, and Arsalan says, "Whoever has executive power might shut out IT from these discussions." So in both cases, we have this impact of AI on senior leaders in an organization. Any thoughts on that?

Art Hu: Let's start with the last one, around leaders who have power shutting IT out. And I think this is a bit of a fallacy in my mind. Or if that's really what's either happening or perceived to be happening, then I think there's something wrong with the relationship between the CIO or the technology function more broadly, let's call it, and that business.

And just very practically, the whole point about technology is to empower the business. So... and this has a little bit with the first question. I would love to be out of my job as it is, because if we can make the technology so democratized, so modular, and so low barrier to entry. If the low code, no code, and agentic framework really comes into its own, and pick your timeframe, I would love to have a very different job, and so fundamentally, I think if there's a business leader that is somehow perceived as shutting out IT, then I think that relationship needs a reset.

'Cause that doesn't even make sense, 'cause it means I want to shut down or... It... There must be something else there. So my first question, if you parachuted me in and I could have a frank conversation with either the technology leader or the business leader or together is, what's causing the need to shut anyone out? Because this is the same thing with shadow IT. Shadow IT is in and of itself not bad. It's a symptom that there's some unmet technology need that the central IT team or the technology team can't provide.

And so it's a bit of an anti-pattern. Anytime I've seen the business feels need to shut out or hide or not say something about what they're doing, it's because there's a bit of an underlying current that's not healthy about the interaction dynamic between the technology and the business functions. They have to be partnered and shutting out just means for whatever reason there's not the trust. And so in that situation, I think the most important thing for that company is to help reset that relationship to start building the trust.

Now leading to the question, the first, the word you started that one, in the compound question about, you know, we don't need a C-suite anymore. Well, it goes back to the earlier one. I think, yeah, I... We may not need the C-suite in the current form. But someone has to lead the companies. I don't think anyone here believes in the next five to 10 years we're gonna have... We are all going to be working just for robots. Because again, this is a... This is a human choice, to the point. Building an enterprise, building a business, building a startup, building something new. These are... And we're gonna get a little bit philosophical here. Those are fundamentally human things. And because at the core of it what... And I'm very open.

If we modify the question, not to say we don't need a C-suite, but if we say, "We don't need the C-suite in its current incarnation." I could totally buy into that. In fact, I hope, I'm certainly working that by increasing the democratization, I hope we can have way more citizen... And citizen developers at Lenovo who are able to contribute, using our platforms, using the agents that we've built and building on top, extending them, doing things that I definitely didn't think of, my team may not have thought about it. That's fundamentally what the platform is for.

And so I wouldn't say we don't need a C-suite because again, words are not so important here. I think the point is we will always need some leadership. We need someone to set the vision for the company. That's fundamentally a human thing. If you believe that that's not human thing, then we have a very fundamental different vision of what the future of technology is. But I'm very open. We don't... The role of the CEO, the CTO, the CFO... Yeah, for sure these are not immune. There's no armor that comes with being in the C-suite. And in fact, the pressure is extremely high to go find the next and what's new.

And just as for our frontline managers, frontline operators, first line support people, their jobs will evolve. There's no exception. The C-suite, the senior leaders absolutely have to evolve. So if you permit me that slight tweak, I agree. We don't need... And in fact it would be weird, if the C-Suite operated the same way in five years and 10 years, after sustained investment and transformation and evolution and bumping up against the frontier of the possible with what these capabilities afford.

Michael Krigsman: It's so interesting to hear your perspective on embracing adaptability, embracing the change, recognizing that... Again, this seems to be the common theme, recognizing that there is change and I don't see you pushing away that change, but rather trying to understand its roots in order to embrace it. And therefore take advantage and essentially roll with it.

Art Hu: One thing for leaders and managers to keep in mind is this is part of, I think, leading with empathy, and leading with authority as well, because it's... I think why run away from it? Yeah. If augmented by an agent I can make decisions better and I can do it, then we should. There's no one who's immune. So I think that's very important. I think it is also... But again, for everyone, again, at any level, I think it really... The hard part and the challenging part is to think through what is as a person, and you know, in Lenovo we sometimes call it carbon-based labor, as opposed to silicon-based labor. What's carbon ba- as a carbon-based, labor, representative, what's my unique contribution? What can I do that is still better?

The Human Element in AI and Chess

Art Hu: And again, this sometimes gets a little bit philosophical. We'll take chess as an example. It's been some time actually, that the best human player in chess who is unassisted, there's no chance for that person to beat the best chess engine. And that's been true for decades now. That's been true for decades. And if you were before, you say, "Well, I guess chess will just die because now it'll just be machines moving the pieces around." But in fact, competitive chess has evolved. There's new forms. It's as popular as ever.

And you would say, "Well, why would people watch..." Admittedly by any objective performance, if objective means ability to win the game. People who are watching chess are watching an objectively inferior competition... and yet, it's as popular as ever. It delights millions of people around the world. Why? And it's because there's still something intrinsically human about that, and I'm an optimist around that. I firmly believe this is why we still need people, representative of carbon-based labor to really think about what are the important things.

And actually, AI accentuates that fundamentally. In a world where answers and data are now easy, it's more important than ever that humans in the loop really mean what are the questions that are worth asking. What are the ones that are worth paying attention to? And so again, I'm an optimist, but I even in the business world, if you say, "Oh, well, the bot said this," people will say, "Why? Does it really understand our business? Has it really that ingrained knowledge to make the call by itself?" And the answer today is no. And so it really forces hard discussions, and it pushes the boundaries of fundamentally what can humans continue to do and add to the enterprise.

Lenovo's Role in AI and Hybrid Cloud

Michael Krigsman: Chris Peterson says, "Lenovo so far isn't a name we usually hear in talks about building out hyperscale AI data centers in the US. What roles are you playing and looking to play in the future in the AI supply chain?"

Art Hu: Many of you know us as making ThinkPads and some of the world's greatest laptops and computing gear. But we also have GPU hardware. We have servers. We have edge equipment. And we also the team and the group that I'm the chief technology and delivery officer for, the solution and services, it's about helping make AI and the hybrid cloud real for companies in ways that are linked for the business.

And so, I think this is one of the points where as computing intensity increases, so if we look at how data centers are being built, the power density is going up by an order of magnitude in terms of how much power is required because of the GPU dense compute, both on training, but ultimately, it's gonna show up in inferencing, as we continue to move forward.

And so for Lenovo, it's win-win. We have the compute that powers the devices because we will always fundamentally at the edge and at the client device level. We'll need phones, we'll need laptops or whatever the next iteration is. Maybe it's more voice led, but people always need some way of interacting and accessing that compute. Because all of this is really compute to help you achieve your outcomes better. Personally, with your friends, at work, all of those.

So whether that comes from GPU compute or CPU compute or more models or more agents, people still need to consume those. And so we're very bullish there about continuing to provide the next generation of hardware at the client end that will help you consume that in the best way, whether it's a display, whether it's augmented reality, whether it's a phone, whether it's a next generation wearable. That is something we're fundamentally bullish around.

And then we move away from the edge and the client, our infrastructure group provides basically the compute that powers the internet. So again, whether it's CPU dense compute or GPU dense compute, we are working across all segments of the market. We work with public cloud providers. We help enterprises with their private cloud and on premise. We're increasingly helping our customers with managing hybrid cloud because large enterprises have a mixture. They'll often have multiple public cloud providers as well as a private cloud.

And so bringing that together is a non-trivial challenge in managing scalability, managing cost, managing transparency, observability, all of those things. And so fundamentally, that's another area where rising demand is something that we're extremely well positioned to meet.

And then finally, landing it for enterprises is very much about services. We're also very aware that... And I'm a CIO. This is where the dual nature and having multiple hats is very helpful for perspective. As a CIO, I'm not excited at all to say, "Oh, let me buy a new server." That does not excite me. But I am very excited if you say, "Well, I've got a solution that will help your legal team manage your multi-billion dollar litigation with much more accuracy and with much less staffing. Oh, and by the way, you have to buy a few servers and a few agents, and we can help you with that." That's an exciting conversation.

And so I think relative to all of those things, because these are fundamentally computing intense, we're gonna need a way on the backend to provide all that computing power. We're gonna need devices and ways, and think about devices not necessarily in the traditional sense, because as I said, who knows what the future form factors will be? But you need some way to consume it. And then you need some way to land that in the enterprises, because enterprises are trying to use this for their business as well.

So, I think that was a broad answer of, along any of those dimensions. But because fundamentally this is going to be a rising tide, and of course there's going to be bumps along the way, but it's going to be around for decades. Think about steam power took decades. Electricity took many decades to really reach. They're going to be here a long time, and Lenovo as providing the things that are underlying all of these. I'm fundamentally very bullish about each of those aspects, and of course, therefore, the whole picture.

AI-nomics and Ethical Considerations

Michael Krigsman: Here's a question from Lisbeth Shaw. "You put out a, Lenovo put out a report that discussed AI-nomics. Can you talk about what is AI-nomics?

Art Hu: AI-nomics is a research report that we did and we released earlier this year that talks about how our enterprises are thinking about agentic around generative AI in 2025. We do one every year. This was the theme we picked, and the main findings is that there are tremendous pressures to demonstrate return on investment, that it's not just a toy. That was number one.

The second part is that... the ability to get the skills ready, we had a question ready about skills, is important because there's still a perceived gap that enterprises are not getting the full value out of this without the right readiness in their workforce. And then finally, one that continues to be common is it's all about the data. No matter how sexy the technology is, no matter how cool the algorithms you use or what frontier model or how many tokens you can get per second on a CPU, if you don't have the right data and you don't use it the right way, also not going to work. So, ROI, skills and readiness, and it's all about the data were some of the key findings that we highlighted.

Michael Krigsman: And this is from Paawan Choudhary, who says, "Thank you, Art, for sharing insights on agility with AI. Could you please share your thoughts on the top three priorities you had when allocating time and effort for various programs during the scaling of AI in the CIO function?" Could do a whole show on this, so three priorities you had when scaling AI on the CIO function.

Art Hu: The first one is the kind of the ROI on paper. Because you want something that you think is good return for the investment. The second one is actually the probability of success, like sometimes we all know business cases look great on paper, but they're actually a bit stickier to realize. So who's actually doing this? Do they have a good track record? So, that's number two.

And then the final one is, I would call it the lighthouse effect. If we are successful, how can we make this a very attractive, package it up so that it creates disproportionate excitement, where we can really get people fired up to say, "Oh, okay, that's what good AI implementation looks like. I want to do something like that." So ROI, probability of success, and then ability to create a lighthouse to excite, other people in demonstrating what's possible.

Michael Krigsman: We have two questions now on the unethical dimensions. One, the first one is from Nicole Jefferies on LinkedIn, who says, "One of the societal concerns about AI is its potential to even further widen the digital divide. How are you thinking about longer term implications to people and societies around the world?" Again, we could do a whole hour-long discussion, days of discussion on this one, so but keep it pretty short.

AI's Role in Addressing Global Challenges

Art Hu: There is a reality around priorities, which is, for example, for the many people in the world who unfortunately today do not have access to reliable electricity or even water and enough food, using AI is not going to be on the top of their concerns. And that's something we don't always talk about enough, but that is a fact. And that needs to be acknowledged. Now that being said, I think the point about AI and what gives me hope is that because it is a general purpose technology, that it is something that we can use, just as we talked about and I spoke about earlier. I want to use it within my company to lower the barriers for access, to lower the barriers. There are many, many entrepreneurs as well who are using AI and applying it to solving the additional social problems of providing better access to what we would consider the basics of living a dignified life. So, I think there, I think the optimism if you look around, there are many social entrepreneurs who are grabbing onto AI to help increase their reach to really helping everyone on the planet.

Michael Krigsman: Another question related to this from Arsalan Khan. He says, "As we have outsourced manufacturing, hired contract software engineers, would we be contracting out AI capabilities to other organizations that might be better at doing the particular task, but who do not necessarily share the same ethical boundaries as our own organization?" So, the question is about outsourcing AI to technically proficient groups who may not share our values and perspectives.

Art Hu: Right. Well, and this one, I think is one instance of a general class of problem, but I think of this typically as, you know, don't outsource your brain. Just, today, if you are, just as when you would use a consultant, all right, you can imagine if you were thinking about any project you would, it's up to you as the project owner or the business owner, you're accountable for the outcome. You may choose someone to be responsible to help deliver it, to execute it. But in terms of, so Arsalan for your question, I think that's part of the due diligence. By making sure, A, you yourself are clear on what it is you are looking for, and then B, as part of selecting who are the partners or the providers with whom you would work, that you do the due diligence on, just like you would on their financials, on their ability to deliver, on their soundness as a company, on their customer references. That, especially if there's reason to believe that it's going to get into some things that are sensitive for you or your company, that you would want to make sure that is a conversation. And so I think the point here is not to walk in blind. You would apply the same level of due diligence as you would in other parts of your business. Yeah, and that's one super important message. There's nothing magical or special about AI where somehow core solid business basics don't apply.

Michael Krigsman: You are evaluating a vendor and they are technically the best. They're great. Be more profitable working with them, but they don't share your values. How do you draw that line and make those distinctions and make the decision?

Art Hu: Well, I think it's the weight. If you just... Very tactically, if you think about a procurement and a vendor evaluation scorecard, that gets a line, that gets a weight. And from a committee of the decision-makers who represent the interests of the company, then you will be able to properly score that and make an outcome. So for, even if it's great on all the dimensions Michael just mentioned, but you, it's really... It's a knockout criteria that, "Boy, we're not sure these people are gonna be ethical," then you could knock them out. And so I think it's making sure then that the point about ethics makes its way into your processes, including about due diligence and procurement.

Michael Krigsman: So you're going through a thoughtful evaluation process to, as you said, to weigh the pros, to weigh the cons, and then you come to the best decision you can come to.

Art Hu: Exactly.

Michael Krigsman: As you talk with Lenovo customers, what are the AI challenges they face?

Michael Krigsman: Where can CIOs add the most value to their organizations in this new AI-based world?

Art Hu: The first is making sure that, one, the executive team is aligned about what is and isn't possible, right? Secondly about really changing the culture by providing the right tools, by holding the right discussions, by creating the right environment, as we said, to invite people in and not use a hammer people over the head to adopt AI, right? And then the final piece I think is role modeling, right? I think there is an expectation as the CIO or- or the leading technology figure at your company to practice as you preach, right? You don't wanna forget about your, your own teams, right? How do you upskill them and make sure they are properly equipped and that they're walking the talk as well?

Michael Krigsman: What advice do you have for CIOs when it comes to scaling AI?

Art Hu: I think three things. One is create friendly competition to get adoption in scale, right? There's nothing better than having the business with you, arm-in-arm, saying that, "Hey, we did the best project ever." Second, stay grounded, right? And that means figuring out where you are as a company, not where you hope you are, but where you are and get the right on-ramps. Don't under or over assume knowledge of where your business is. Uh, and then finally, I think sometimes people call it a race, but that's not quite right. So the right mental model is that it's more of an ongoing journey. Definitely not a sprint, not even a marathon because you don't finish, right? And recognizing that each company has their own journey to be on, not... And that you're not necessarily r- running the same race as someone else, right? And that's really the sweet spot more than ever about sitting at the intersection of really a tremendously set of dino- uh, tremendously dynamic set of technology, business, and market forces that I think CIOs can really help bring their companies to the next level with that recognition.

Michael Krigsman: With that, we're out of time. A huge thank you Art Hu, he's the Global CIO at Lenovo. He's also the Chief Delivery & Technology Officer of Lenovo Services and Solutions Group. Art, I- I'm so grateful for you to have come back and spent your time with us. Thank you so much.

Art Hu: No, thank you, Michael. You made it easy with your insightful questions. And thank you to the audience for your great questions also.

Michael Krigsman: Go to cxotalk.com and subscribe to our newsletter. And we want you to come back, audience, to our upcoming shows. We really do have tremendous shows. This video will be posted on the CXOTalk website by Monday. Tell other CIOs you know to come check it out. We'll have a summary. Th- this, there's, there's a lot of valuable information for chief information officers here, so I encourage you, share this with the CIOs you know because there's a, there's a lot here to learn. Thank you so much, everybody. Thanks to Art Hu, and we'll see you again next time. Take care.