Understanding sovereign AI: What boards need to know now

Summary

---

The Corporate Director Podcast

Episode: Understanding Sovereign AI: What Boards Need to Know Now
Date: February 24, 2026
Host(s): Dottie Schindlinger (Executive Director, Diligent Institute), Megan Day (Strategy Leader, Diligent)
Guest: Josh Klein (Board member, CEO, AI strategist, advisor to countries and Fortune 500 companies)

---

Episode Overview

This episode delves deeply into “sovereign AI”—the role of artificial intelligence as core national infrastructure—and what that means for corporate boards. Through a thought-provoking conversation with Josh Klein, a seasoned AI strategist and advisor to both corporations and governments, the episode explores how rapidly accelerating AI adoption impacts governance, risk management, board strategy, national policy, and the future of business leadership. Practical insights and candid dialogue focus on how directors and executives should be engaging with AI—not at a surface level, but with real curiosity, strategic thinking, and hands-on experimentation.

---

Key Discussion Points & Insights

1. State of AI in the Enterprise (00:54 – 07:17)

• Explosive Growth:
  • Organizations have increased workforce access to sanctioned AI tools by 50% in just a year, with ~60% of employees now using approved AI in their workflows.

    “They've said that companies have increased workforce Access to sanction AI tools by 50% in a single year.” – Megan Day [01:36]

• Nature of AI Adoption:
  • A minority (34%) of companies are using AI for deep transformation; most are focused on productivity and incremental efficiency.

    “AI is delivering productivity for most, but business reimagination for few.” – Dottie Schindlinger [02:09]

• Board Strategy Disconnect:
  • Directors find overseeing strategy their biggest challenge amidst AI-fueled volatility, reflecting a gap between AI’s promise and actionable business vision.

    “We still don't have a clear picture of what that means for this company… we're probably not really digging in the way that we should to figure out, you know, how we could change, not just little incremental change, but big sweeping change.” – Dottie [03:30]

2. Governance & Risk Oversight Shifts (04:54 – 07:17)

• Rise of Risk Committees:
  • Board audit committees are becoming overwhelmed with oversight across all domains, prompting the creation of specialized risk or innovation committees.

    “Their remit was just getting ridiculous...Is it really being effective?” – Dottie [05:33]

• Acceleration from Experimentation to Enterprise Impact:
  • As AI moves from pilots to core business operations, directors must wrestle with more complex, systemic issues.

    “Lots happening, lots accelerating. And that means lots for directors to wrestle with as this sort of moves from experimentation to enterprise wide impact.” – Megan [06:50]

3. The Concept of Sovereign AI (10:14 – 13:41)

• AI as National Infrastructure:
  • Governments are beginning to treat AI as vital infrastructure—shaping “soft power” by ensuring their values and priorities are embedded in national AI systems.

    “If you're the average of your five closest friends...AI is going to be one of them. And so it behooves governments to have a voice in what that one is.” – Josh Klein [10:14]

• Beyond Efficiency:
  • Unlike previous tech waves like cloud (which extended existing processes), AI is a “memetic engine”—a system that mimics human communication and thinking.

    “Now we have AI, which in essence is a memetic engine. That's not anything we've ever had before.” – Josh [11:13]

• Predominant Role as Soft Skills & Companionship:
  • Most people use AI for advice, therapy, and social interaction, not just work tasks.

    “The predominant use of AI today is companionship...it's people looking for therapy, advice on what to do with their boyfriend, you know, how to cook a lasagna.” – Josh [11:49]

4. AI's Double-Edged Sword: Promise & Risks (12:56 – 13:02)

• AI Passes the Turing Test—and Is Weaponized:
  • Most phishing (cyber fraud) attacks are now AI-generated, and they are more effective than human-written ones.

    “The majority of phishing attacks...are being produced by AI. And generally speaking, those attacks are 60% more effective than human written ones.” – Josh [12:12]

• Should Boards Be Excited or Terrified?
  • “Well, you should be terrified.” – Josh [13:00]
  • But also excited: This is a profound, historic shift; AI can augment and amplify uniquely human capabilities.

5. Industry’s Role & Responsibilities (13:41 – 15:16)

• Faster than Government, with More Responsibility:
  • Industry can—and should—lead in upskilling, innovating, and shaping ethical AI, not just waiting for government regulation.

    “If we've learned anything over the last five years, now more than ever, companies need to own that role and have that voice or they get obviated.” – Josh [14:31]

• Advice to Leaders: Management Over Technical Mastery
  • Engaging with AI is “essentially a management challenge,” rooted in decision-making, contextualization, and communication.

    “Working with AI is essentially a management challenge. And the better it gets, the more so that is the case.” – Josh [15:16]

6. AI in Action: The Iceland Example (16:27 – 18:56)

• Real-World Uptake:
  • In Iceland, 85-90% of students use AI weekly, and 40-50% use it daily (often for homework and even dating advice).

    “Somewhere between 85 and 90% of all kids in the country are using AI weekly and somewhere between 40 and 50% are using it daily.” – Josh [16:46]

• Game-Changer in Education:
  • AI as an individualized “tutor for every student” could propel national performance, if systems adapt to support it.

7. Risks: Hallucinations, Accuracy, and Human Loops (18:33 – 19:45)

• Ongoing Concern:
  • Models are improving, but AI outputs can be wrong. Boards should calibrate expectations—technology doesn’t exceed human fallibility, except perhaps in life-or-death fields.

    “Have you ever worked with a human being that didn’t make mistakes?...Why? You know, if I could hire someone who had encyclopedia knowledge...if it was wrong, occasionally, I don’t think I’d throw out my hands.” – Josh [19:00]

8. What's Next: Alignment, Agentic AI, and Continuity (20:12 – 21:56)

• Model Alignment & Ethics:
  • Boards will increasingly need to grapple with “alignment”—how AI models reflect societal values, avoid harm, and align with company purpose.

    “What does it mean for an AI model to support Western democratic values...these, again, memetic engines have a huge influence on human beings.” – Josh [20:12]

• Ongoing, "Agentic" Relationships with AI:
  • Next evolution: AI agents with persistent memory, ongoing relationships, and deep integration with user workflows (calendar, documents, etc.)

    “It has one permanent ongoing conversation with you and a memory framework. So it has compounding and iterative memory.” – Josh [20:44]

9. Risks of Workforce Churn, Talent Gaps & Leadership Pipelines (22:09 – 26:33)

• Death by Churn:
  • Companies reducing headcount thanks to AI may be disabling their future talent pipeline and capacity for innovation.

    “The people that you need most for your company to continue to exist a year from now have all been leaving the building.” – Josh [22:09]

• Disrupted Paths to Leadership:
  • Traditional “start at the bottom, learn, rise to executive” paths may vanish; directors must rethink how to equip future leaders in an AI-dominated environment.

    “That pipeline is getting disrupted and how do we cultivate leadership if they're not learning the basic ropes?” – Dottie [23:24]

10. Boardroom Advice: Questions and Culture, not Technical Certification (26:33 – 28:54)

• Key Questions for Management:
  • What tools are people using? For what? How does culture support AI exploration?

    “If you get blank stares, then that's where you got to start.” – Josh [26:52]

• Directors Don’t Need to Be Tech Experts:
  • The right questions and curiosity matter more than technical depth; subject expertise + AI tools yield results.

    “You don't need the technical capacity to explore where there can be value. All you need is access to an AI and an ability to ask intelligent questions.” – Josh [27:59]

11. Expertise: Redefining the "AI Expert" (28:54 – 31:46)

• No One Is Truly an “AI Expert” Now:
  • The field is too new and too dynamic for a static definition; experience, curiosity, and contextual understanding matter most.

    “Does it make sense to ask whether companies have AI expertise? Sure...But is it worth going out and buying a grad from a Ivy League school so you can check that box? I mean that's a signaling exercise.” – Josh [29:24]

• Evaluating "Experts":
  • Boards should focus on advisors who understand their business and add value, not just technobabble.

    “See if they have an understanding of their business and can add some value. And that doesn't mean spouting a lot of, you know, acronyms.” – Josh [30:57]

12. Action Steps for Directors (31:46 – 32:55)

• Start:
  • Get hands-on; subscribe to an AI model, experiment, learn through use.

    “Start by getting a subscription to a model and asking it what they should know and have some conversations with it.” – Josh [31:59]

• Stop:
  • Don’t just buy “AI for X” packaged tools; work with raw models to better understand potential and risks.

    “Stop buying, how would I call it? Halfway models...figure out how to use the raw tools.” – Josh [32:12]

• Continue:
  • Stay curious, keep asking questions—nobody has all the answers in this fast-changing landscape.

---

Notable Quotes & Memorable Moments

• On AI's pace:

  “It really is happening quickly. It really is pretty...leaves you a little breathless.” – Dottie Schindlinger [02:09]

• On AI’s role in everyday life:

  “AI is going to be one of [your top five closest friends]. And so it behooves governments to have a voice in what that one is.” – Josh Klein [10:14]

• On risk committee overload:

  "Their remit was just getting ridiculous. I mean, they were literally being asked to oversee every kind of risk..." – Dottie [05:33]

• On “death by churn”:

  “The people that you need most for your company to continue to exist a year from now have all been leaving the building.” – Josh [22:09]

• On what directors should do:

  “Working with AI is essentially a management challenge. And the better it gets, the more so that is the case.” – Josh [15:16]

• On expertise:

  “Does it make sense to ask whether companies have AI expertise? Sure...But is it worth going out and buying a grad from a Ivy League school so you can check that box? I mean that's a signaling exercise.” – Josh [29:24]

• On the role of humans in the future of governance:

  “I think the biggest difference [in boardrooms ten years from now] will be the number of humans in the room. Just to throw something spicy in there.” – Josh [33:08]

---

Important Timestamps

• 00:54 – 07:17: State of AI in the Enterprise & Board Oversight Challenges
• 08:45 – 13:41: Josh Klein on Sovereign AI, AI as Infrastructure
• 16:27 – 19:45: Iceland & AI in Education, Hallucination Risks
• 20:12 – 21:56: Alignment, Agentic AI, Ongoing AI Relationships
• 22:09 – 26:33: Death by Churn, Leadership Pipeline Risks
• 26:33 – 28:54: What Boards Should Be Asking
• 28:54 – 31:46: Rethinking the Idea of “AI Expertise”
• 31:46 – 32:55: Josh’s Start/Stop/Continue for Directors
• 33:08 – 34:41: Lightning Questions; Flat Organizations, Human Roles
• 37:53 – 43:58: Hosts’ Analysis on AI Expert Paradox, Worker Experience Loss, Innovation Risks

---

Related Reflections & Key Takeaways

• The term “AI expert” is now nearly meaningless; real board value comes from curiosity, pattern recognition, and domain understanding, not credentials.
• Directors should get hands-on with AI to build intuition and informed skepticism.
• Companies need to balance the drive for efficiency with sustained human-led innovation and talent development. Over-reliance on AI without cultivating future leaders is a strategic risk.
• Governance frameworks (risk, innovation, audit committees) should evolve to match the unique challenges and pace of AI disruption.
• Industry has a critical role not just in adoption but in setting norms, ethics, and responsible guardrails in advance of slower-moving government policy.

---

Final Thoughts

This episode is a sharp, candid, and practical exploration of how directors and executives should approach AI—not with fear or awe, but with hands-on learning and open, contextual questioning. Boards must stay strategically engaged, embrace ambiguity, and foster a culture of experimentation—not just to survive AI’s disruption, but to harness its transformative potential.

---

Transcript

[Megan Day]

Foreign.

[Josh Klein]

Welcome to the Corporate Director Podcast where we discuss the experiences and ideas behind what's working in corporate board governance in our digital tech fueled world. Here you'll discover new insights from corporate leaders and governance researchers with compelling stories about corporate governance strategy, board culture, risk management, digital transformation and more.

[Dottie Schindlinger]

Hi everybody and welcome back to the Corporate Director Podcast, the voice of modern governance. My name is Dottie Schindlinger, Executive Director of the Diligent Institute and I'm joined once again by my co host extraordinaire Megan Day, strategy leader here at Diligent. Megan, how are you doing on this slightly warmer day?

[Megan Day]

Thawing out Dottie and using my newly defrosted fingers to dig into something I would actually your take on. Deloitte just released their 2026 State of AI in the enterprise report and I think it paints a really vivid picture of where organizations are headed with AI. Our favorite topic.

[Dottie Schindlinger]

Our favorite topic. Well, and Megan, you sent this over to me and I got a chance to skim through it. There's a lot in this report. In fact, I just took the time to read the executive summary and there's a ton in just the executive summary. But what were some of the things that really stood out for you as the big takeaw ways? I think I know what my favorites are.

[Megan Day]

Yeah, for me it was really some of the quantification around the acceleration that I, I feel like I am feeling personally. They've said that companies have increased workforce Access to sanction AI tools by 50% in a single year. And that means that now around 60% of employees have approved AI in their workflow. Does that pace surprise you? I feel like it doesn't surprise me, but it is kind of jarring to think of the change that has happened so quickly.

[Dottie Schindlinger]

Does it surprise me? No. I mean, I think we all knew coming into this year that every company was making big bets on AI. And you know, and the way that most companies are approaching making bets on AI right now are around productivity gains. Right. So how can we build efficiency? How can we build productivity? And in fact, that's another stat in this report, that AI is delivering productivity for most, but business reimagination for few. So they're saying 34% of companies report they're using AI to deeply transform their business. That's a minority versus those that are using it to just, you know, speed things up, make things cheaper, make things faster. It doesn't surprise me, but I agree with you that even though it doesn't surprise me, the speed with which this is Changing is still pretty awe inspiring. It's still pretty. It leaves you a little breathless.

[Megan Day]

Yeah.

[Dottie Schindlinger]

When you see it all spelled out on a page, you're like, yeah, you knew it was fast, you knew it was happening, but it really is happening quickly. It really is pretty, you know, kind of makes your head spin around.

[Megan Day]

Now, Donnie, I would love your take your comment about the reimagination piece versus maybe some of the more surface optimization that many organizations have embraced. You know, do you think boards are pushing hard enough in terms of like, let's do blank page re envisioning of what this organization means?

[Dottie Schindlinger]

What a good question, Megan. Well, let's see. I mean, what I will tell you is I think boards would tell you they're not pushing hard enough. I mean, one of the things we heard in what directors think that, you know, we've obviously talked about quite a few times on the show already this past year, was that directors felt that the area of the business they're having the most challenge overseeing right now is the strategy. And we, we thought, wow, that's a really interesting answer given that strategy is such a bread and butter governance issue. But I think that's what they're saying. I think what they're saying is, look, you know, you guys have been talking about AI and making big bets in AI. We still don't have a clear picture of what that means for this company. And we also don't feel like we really fully have our arms around what the strategy is here. And so I think directors themselves would tell you, yeah, we're probably not pushing hard enough. We're probably not really digging in the way that we should to figure out, you know, how we could change, not just little incremental change, but big sweeping change. You know, it's an interesting time to try to have some of those conversations because there's so much happening in the broader world that is so volatile and uncertain and complex and ambiguous. And so for this to be a time to reimagine your business might feel a little weird. It's like, I feel like a lot of companies are just trying to figure out how to stay alive, you know, with everything being so upside down. I don't know, I don't know what you think about that, Megan, but I. Those are the two things that sort of occurred to me when you asked that question.

[Megan Day]

Yeah, I mean, I'm always one that gets excited by the art of the possible. But I think your call out around the risks and uncertainty is very true. And I was actually just reading something this week about the rise of the risk committee in in particular U.S. organizations. What was once sort of relegated to financial services and those big organizations has now become increasingly more popular as a result of trying to wrap their heads around all of this complexity. But I do think it is an interesting question, an interesting ask, and one that you know, should at least come up in, in conversation at the board level.

[Dottie Schindlinger]

Well, I love that comment, Megan, because I feel like we've talked about this a few times that, you know, the, the remit of the audit committee, which if you don't have a dedicated risk committee, tends to be the place where risk lives. I mean, their remit was just getting ridiculous. I mean, they were just, they were literally being asked to oversee every kind of risk, no matter what was involved, whether it was, you know, AI risk, cyber risk, financial risk, climate risk. Like they were just getting asked to oversee every kind of risk. And it is I think, a time to start to think about, okay, is this, is this still the right balance? You know, there was something to be said about putting kind of the audit committee's type of discipline onto areas of risk, but after a while it becomes so bloated and there's so many different things they're having to keep their eye on. Is it really being effective? So maybe separating out a separate risk committee or you know, a technology and innovation committee that's just purpose built, focused on business transformation in light of AI, you know, how are we thinking about cyber? Those things could be really, really important depending on your industry, depending on how quickly you're moving, depending on what's happening. I'm not surprised, I'm not surprised that we're starting to see more companies adding a committee either called Risk or Innovation or of that sort.

[Megan Day]

Well, lots happening, lots accelerating. And that means lots for directors to wrestle with as this sort of moves from experimentation to enterprise wide impact. And I think, Dottie, this report, if you haven't had a chance, our listeners to check it out, please do so. And I think it tees up nicely the conversation you had with a gentleman who just sort of blew my mind in the conversation that we're about to listen to.

[Dottie Schindlinger]

Yeah. So Megan, perfect segue. I was so excited to have Josh Klein back on the show. Josh is, he is really just one of the most thoughtful and interesting and quite frankly, I think smartest people I've spoken to in AI. And even he would say, I don't even think of myself as an AI expert. Meanwhile, that's all he does for a living. He's being hired by countries to help them think about sovereign AI. So, you know, I would think of him as an AI expert, but it was great to have a chance to sit down with him and really get below the surface. I feel like I've had just way too many conversations about AI this year that stay at kind of a high surface level and don't go deep enough. And I think that's a real challenge for a lot of corporate leaders right now is how do we get below just the. Yeah, AI. Big deal. Scary, exciting. Next. You know, you need to get below that. We need to go deeper than that. So it was great to have this conversation with Josh. I think you're gonna enjoy listening to this one. There was a lot. We unpacked, a lot of different issues. But let's give it a listen, then come back and talk about some of the things you heard. Joining us on the corporate director podcast today is Josh Klein. Josh Klein is a board member, CEO, and strategist with a proven track record of advising countries and Fortune 500 companies on navigating disruptive technologies like artificial intelligence. Josh, thank you so much for joining us on the show again. It's great to have you.

[Josh Klein]

Absolutely. My pleasure.

[Dottie Schindlinger]

Well, I just gave the most cursory explanation of all that you do, so let's start off by having you tell our audience a little bit more about your career background and some of your current gigs.

[Josh Klein]

Sure. So, mostly through luck, I had the good fortune to stumble through a number of emerging technologies, got to work on the first e commerce engine, the one that VeriSign bought. I was in the telco business with the launch of short codes, Misbegotten as that was, was in cloud deployments, with one of the first deployments of medical data. So I got to see a lot of new and emerging technologies kind of happen and disrupt the markets. Eventually, I ended up where I am now, which is mostly working as an advisor, consultant, and then CEO on a number of different initiatives nowadays almost entirely about AI.

[Dottie Schindlinger]

Okay, I have a. I have a side question to ask you. How big of an AI workforce do you manage these days?

[Josh Klein]

I think right Now I've got 12 agents that run continuously.

[Dottie Schindlinger]

That's awesome. Well, listen, one of the reasons I wanted to have you back on the show is you're doing something really interesting that I don't think a whole lot of corporate directors are paying attention to, and I thought it would be great for us to talk about that, and that is that you've been doing a lot of work lately helping different countries, including Iceland. And Zanzibar and I know there's other things sort of in the works. Create their AI development policy at the national level. I wondered if you could tell us a little bit more about this process. What does that mean? What are some of the things that go into it? I'd just love to hear a bit more about that.

[Josh Klein]

Well, it's interesting because, at least initially, a lot of the conversation around, you know, sovereign AI, if that's what you'd like to call it, followed the corporate tracks, right? Like governments need to have efficiencies, they need to see growth. You know, all the, all the same drivers that companies would have. And there's a lot of inefficiencies in most governments. But over time, I think a lot of different countries started to realize that beyond the benefit that a standard new technology would represent along those lines, there's also the fact that AI really is intense soft power. If you think about it this way, or rather the way I explained it to my mom, is if you're the average of your five closest friends, if your behavior and how you live life is the average AI is going to be one of them. And so it behooves governments to have a voice in what that one is.

[Dottie Schindlinger]

So I'm going to say it in a way that probably makes no sense, but this is how I think about it. Is it basically equivalent to countries thinking of AI as infrastructure?

[Josh Klein]

It's absolutely infrastructure. But my point is that it goes well beyond just making the lines of the DMV shorter, right? Like that's, that's easy to understand. I think what's, what people are starting to grasp is that unlike cloud, for example, which basically took your office suite and made it so that you could collaborate with others more easily and have redundant backups and stuff like that, that basically extended what the technology did before. Now we have AI, which in essence is a memetic engine. That's not anything we've ever had before. So as a result, the predominant use of AI today is companionship. It's not actually to improve workflows, it's companionship. It's people looking for therapy, advice on what to do with their boyfriend, you know, how to cook a lasagna. Like, it's very much a soft skill relationship machine, which is a different beast.

[Dottie Schindlinger]

So that's super helpful. Although I need you to explain mimetic engine for people like me who don't understand what that means.

[Josh Klein]

Sorry, sorry. So the concept of memetics being that human beings in many ways think in language the Ways that we construct our arguments, our worldview, the way that we understand ideas or topics. It turns out that AI does this really well. We passed the Turing test a long time ago. If I put you online and you chat with an AI or chat with a human, you're actually more likely to think the AI is human than the human being. So right now, you know, as just as one proxy, the majority of phishing attacks. So those are those emails you get where people are trying to trick into doing things. The majority of them, I think 86% of them are being produced by AI. And generally speaking, those attacks are 60% more effective than human written ones. And that number is going up.

[Dottie Schindlinger]

So convince me I should be excited and not terrified of this.

[Josh Klein]

Well, you should be terrified.

[Dottie Schindlinger]

I should be both.

[Josh Klein]

Yeah, both, both. I mean, be excited because this is a profoundly transformative moment in human history. It's really shifting from getting human beings to act like machines to empowering human beings to do what they do uniquely well, better than ever before.

[Dottie Schindlinger]

Okay, so that was super helpful. What should be the role of industry in this process? So you've got countries that are thinking about this on this broad scale basis, really, I think vis a vis their populations. Right. Thinking about how do we roll this out for the country. I know that's going to have an impact on companies. How should industry be working on this effort together or what should it mean for them? How should they think about this?

[Josh Klein]

Well, I think there's two aspects of it. One is to recognize that industry moves faster than government by a big margin, and that's not news to anybody. But it does mean that companies have both an opportunity and a responsibility. The opportunity is if you can provide the skilling opportunities, the innovation opportunities, the cultural transformation to support using these new tools, you have significant market advantage ahead of you. Because not everyone is going to move quickly on that. In terms of responsibility, society moves slowly, particularly at a governmental perspective. So if you are willing to take a role in which you have a stand on these issues, you have a chance for your voice to be heard. And if we've learned anything over the last five years, now more than ever, companies need to own that role and have that voice or they get obviated.

[Dottie Schindlinger]

So, you know, obviously companies move faster than countries. Technology moves a lot faster than either. So, you know, I'm just thinking about directors and I'm thinking about senior management teams and this feeling that I hear every time I talk to them that how in the world do we keep up with this? There's such concern I think broad based concern among everyone that a lot of people who are running companies don't really have the deep knowledge that they need in this technology or really any technology, let's face it, to be able to kind of keep up with the change and to keep up with the level of decision making they need to make on the fly. Do you have any suggestions? I mean, you've been in this world for so long, to you this is like breathing. But do you have any suggestions for those of us who aren't you, Josh Glein what do we do?

[Josh Klein]

Yeah, but I mean it's the same thing I did. You just have to go and play with it. And the good news is, particularly if you are a leader, you are more well suited to this than to almost any other technological challenge that has come. Because working with AI is essentially a management challenge. And the better it gets, the more so that is the case. There's an analyst by the name of Balaji that says that AI is really good at middle to middle. So in other words, if you have a challenge, you need to contextualize it, you need to situate it within the operating environment. You need to decide what the key questions and the constraints are and you need to provide rules for engagement for the AI to figure out how to solve those problems. Then AI can go through and do a lot of the middle part. But at the end you're going to have a solution that then needs to be recontextualized into the real world re situated, relevant to your stakeholders and the people whom it will affect recommunicated to other human beings.

[Megan Day]

Beings.

[Josh Klein]

And that also requires human beings. So that middle to middle piece is growing and the capacity is getting better and better. But those two other pieces are pieces that, you know, human beings, particularly leaders, have owned for a long time.

[Dottie Schindlinger]

That's really interesting. Let's go back to talk a little bit more about some of the work that you're doing with the countries. Let's just start with Iceland because I know that's that's where you currently live. And so talk to us a little bit about, you know, maybe go out a couple of years from now. What's going to look different, like what is going to be different when your work with Iceland is, is in place?

[Josh Klein]

I mean, would that I knew for sure. I can tell you what I think the potential is, which is right now there's a bunch of different trends happening with AI and again, it is moving incredibly quickly. In fact, it's funny, I talked to an old friend of mine Just last week. And I said, did you hear about such and such? He's like, no, no, I'm so far behind. I'm like, I know that was five days ago. So like, that seems like a long time ago. So from a governmental perspective, I think we have a couple things colliding. One is that consumers or citizens are absorbing AI at an incredible rate. So like, I've been working a lot with the education Ministry here. One of the things that we've seen is something between 85 and 90% of all kids in the country are using AI weekly and somewhere between 40 and 50% are using it daily. Now when I say they're using it, of course they're using it with homework, because wouldn't you. But also like I spoke to my 13 year old niece and said, what's the main thing you use AI for? And she said, dating advice. So like people are using this and they're using it in the way that suits them most. So that's one thing. The opportunity on the other hand. And again, just to use an example from education, there's something that they call the, I think they call it the two sigma problem. But in any case, the point is that if you provide a student with a private tutorial, then you can almost guarantee that they will be in the 95th percentile of performance for the class if they have unlimited access to a private tutor. The problem with that, of course, is that you couldn't afford to have a private tutor for everyone, but now you can. So the opportunity is also similarly massive. If we are able to move our constructs fast enough, we can jettison ahead of everyone else in terms of just purely academic. And then of course you can extrapolate from there.

[Dottie Schindlinger]

How much do some of the risks that we hear about AI all the time play into this? Right, because everyone's concerned about hallucinations and about, you know, just false information be presented as truth. I mean, are we getting to a point where that is less of a concern because things have gotten better, the, you know, the engines have gotten better, or is that still something that we need to make sure we have a human in the loop? And I'm thinking if that human in the loop is a 7 year old child that doesn't know the difference.

[Josh Klein]

Yes.

[Dottie Schindlinger]

So tell us what you think about that.

[Josh Klein]

So it's so funny because. So yes, on the one hand, yes, models have gotten much better around hallucination. A lot of frameworks have come out to help ameliorate it, workflows are improving to help Mitigate it, et cetera. But it does happen. It's just because AI is basically a token predictor, right? It's just guessing at what the next word should be. So it's imperfect. On the other hand, have you ever worked with a human being that didn't make mistakes? And why? You know, if I could hire someone who had encyclopedia knowledge of everything that's ever been published on the Internet, and a lot more besides an incredible analytic capacity, and could do all the things that AI could do if it was wrong, occasionally, I don't think I'd throw out my hands and say, oh, this is garbage.

[Dottie Schindlinger]

That's such a good point. You know, I've never really thought about it that way, but you're so right. Like, we demand from our technology things that we don't get from human beings, except for maybe we pilots and doctors that we. We also demanded of pilots and doctors. But other than that, so. So you're a great person for me to ask this question. What are some of the things that you're most excited about related to, you know, recent current innovation or things that you see coming in the near future? What are some of the things that excite you the most?

[Josh Klein]

So one thing that's farther, that's farther reaching, and one thing that's immediate, the farther reaching thing is I feel like we're starting to have more serious questions around alignment. What does it mean for an AI model to support Western democratic values, for an AI model to not accidentally encourage suicide? Like, you know, how do, how are we actually addressing the fact that these, again, memetic engines have a huge influence on human beings because they act startlingly human?

[Dottie Schindlinger]

Right.

[Josh Klein]

How do we manage that? So that's one thing that I'm really excited to see people starting to take more seriously. And a lot of companies, Anthropic is a good one, have published a lot of papers on this kind of research. The other thing, more short term, is just in the last couple weeks. OpenClaw is a software package that has exploded where people are essentially taking an AI agent, where you don't have to start over with every conversation. It has one permanent ongoing conversation with you and a memory framework. So it has compounding and iterative memory. And that's super exciting. Now. It was kind of a dumpster fire at first because it wasn't well controlled. There were some real security issues, but those are getting hammered out. And I think it's kind of shifting us towards the next level of AI agentic use, which is to say, why don't we have an ongoing and permanent relationship with the AI that knows us individually extremely well and has access to our tools. So that could be Google worksuite or Microsoft Office. That could be my calendar, my email, it can go and schedule appointments for me. Like it can do a bunch of these things. And then again it has compounding value by retaining memory. So I think that's another big shift in AI, especially from an executive or a board director view that has big implications.

[Dottie Schindlinger]

All right, so let's also talk about the flip side of that. Now. You kind of touched on some of the big risks, but what are some of the things that concern you the most that you'd really like leaders to hear that they need to get their arms around and do a better job overseeing?

[Josh Klein]

I think the biggest thing is death by churning. So a lot of companies that I'm talking to are sort of quietly celebrating that their headcount is going down. But they're keeping, you know, they're keeping the, the output that they've always had and they're kind of just quietly not rehiring. And in the meantime AI tools and related there's kind of picking up the slack. Well, this is great, right? You've lower operating costs and similar outputs. Total win. The problem is the people that you need most for your company to continue to exist a year from now have all been leaving the building. One of the things that I've been seeing over and over again is that the frontier of AI is jagged. It's really good at some things and terrible at others. And where you discover transformative value is impossible to predict. You need to give access to the tools to basically your entire staff. Try and give as much access to your data as you can and figure, let them figure out where in your workflows they can create that transformative value that does not equal letting the most enterprising entrepreneurial people lead first. So because it's behavioral rather than structural, you don't get a degree that says good. With AI, generally speaking, it's a big risk for companies. So I think that's one of the things that kind of keeps me up at night.

[Dottie Schindlinger]

Yeah, it's a really good point. In fact, I recall a conversation that I had with a group of directors last summer was under Chatham House rule, so I can't share specific details. But what I can say is one of the conversations was about this very issue of if we are replacing human beings with AI, how do we make sure we have a talent pipeline? Because we have certain roles that we've always had people come in, learn at the basic level, get really good, and then we promote them and eventually they make amazing leaders at our company. That pipeline is getting disrupted and how do we cultivate leadership if they're not learning the basic ropes? And I wonder if you have any thoughts about that. I mean, I think it is something we're going to have to grapple with. I don't know that we are grappling with it very well, but do you have any thoughts about how we, you know, how do we get away from disrupting that pipeline?

[Josh Klein]

I think a lot of it has to do with you need to create opportunity, but you need to do it while you're enduring disruption. So one of the things that, you know, I mentioned earlier, openclaw, one of the things I've been doing in playing with it is doing security audits of all the ways that I can give something like that access. And one of the things that I've come across is that there are an immense number of CFO skills, MCP servers, like just frameworks, all these different tools that you can drop in to do a lot of the things that a CFO does. This is true for CMOs and all the other roles as well. I don't mean to pick on CFOs, but the point is that if I need that skill set, I can get 60 to 80% of it as a drop into an AI agent. What's the other 20%? And how is that high value, particularly as AI starts to eat away with the rest of those workflows? And to your point, most of that has to do with executive decision making and how do you cultivate that? Well, it doesn't necessarily mean you have to be doing CFO work. It's just doing executive work.

[Dottie Schindlinger]

Yeah, but learning that executive work, at least for the past couple hundred years, has required learning more basic skills first and then applying them in more complicated situations. So it'll be interesting to find out how do we, how do we shift? You know, I mean, do you have any thoughts about the best way to leverage AI to shift us to get to that executive level decision making, maybe differently or faster.

[Josh Klein]

I think a lot of it has to do with opportunity to take risk. Like where are you creating? And I hate to use the phrase Darwinian, but you know, it is, it is hopefully meritocratic. Where do we create opportunities for that sort of engagement to happen at various levels in the company? Again, you don't learn how to manage a billion dollar P and L by playing in the break room, right? Like Again, you do need to step up and get some real engagements under your belt. But I think a lot of it has to do with how do you use mental models, how do you make good decision making using frameworks, how do you generate buy in? These are all the kind of higher level managerial leadership skills that, that we're all sort of familiar with, but that you get or have gotten in the past, mostly through trial and error or through experience. How can you expedite that in a smaller scale so that the best candidates are able to move faster?

[Dottie Schindlinger]

All right, so let's shift gears a little bit, Josh, because I know you've served on boards and I know you've spoken to lots and lots of companies and board members. So put yourself in the board seat. What would be some of the first few questions you'd want to ask management about the company's AI strategy? I mean, let's say you're new, you're trying to get your arms around it. What are some of the things that you want to ask?

[Josh Klein]

The first thing I always like to ask is what tools are people using and what are they using them for? And if you get blank stares, then that's where you got to start. AI is an incredibly powerful tool. There are also significant risks. If you have no familiarity, you're driving blind. So that's step one. And then beyond that, usually I start drilling down into culture questions again because it's really difficult to predict where you're going to find value, value within your organization. It's easy to say, okay, you're a bank, let's start looking at, you know, what's your deal flow look like? Where you know, what pipelines do you have internally? Where is there greater risk? We can reduce, efficiencies we can reduce. But really, because the technology is so disruptive, it often starts with that behavioral perspective, like, how is the organization supporting the key behaviors you're going to need to find that transformative value?

[Dottie Schindlinger]

Okay, what I love about those questions, none of them are technical. These are just good governance questions. Is that, is that a lesson that directors should take from this conversation, that you don't need to be a technological expert to ask the right questions when it comes to AI?

[Josh Klein]

100%. Yes. In fact, we're at a bit of a transformative moment now. Over the holiday break or the winter holidays, there was a big phase shift in a lot of the big models ability to produce functional code. Now, I'm not suggesting that your vibe coding project over the weekend should be released over the enterprise, but the Ability to spin up an MVP or a prototype that you can actually bang against customers. It's now minutes, not months. So from that perspective, you don't need the technical capacity to explore where there can be value. All you need is access to an AI and an ability to ask intelligent questions. So it really is a huge advantage to leaders where their subject matter expertise, their understanding of the cultural context, their ability to use that executive level reasoning, it can be applied directly today at scale.

[Dottie Schindlinger]

Okay, so we are though hearing institutional investors like BlackRock and some of the proxy advisors like Glass Lewis asking companies to include in their proxy for 2026 any AI expertise that exists on the board. So, thoughts about that? I mean, how should they approach that? I mean, you know, obviously being an AI expert is a very different thing from, you know, taking a couple classes and learning a little bit and playing with it on the weekends and doing some vibe coding. So what, what are your thoughts there in terms of AI expertise on boards?

[Josh Klein]

I mean this is a technology that's been around for what, three years?

[Dottie Schindlinger]

Yeah, exactly.

[Josh Klein]

So AI expert. What like I don't consider myself an AI expert. What, what who's an expert like to get? To give you an example, I have a friend of mine who's a marketer. He had one a CMO role in a mid sized company, but otherwise that was about it. And I don't know anyone that knows AI models better in terms of the marketing space. And that's all nights and weekends that he's put that together. It's just he's been extremely curious and incredibly interested. Another example, I have a friend of mine who's a director who has one of the most finely tuned assemblages of AI models for video and so story production of anyone I've ever met. Is he an AI expert? Like it seems like if I have any questions around that domain that he's the one I'm going to ask. But that's not what's on his business card. So does it make sense to ask whether companies have AI expertise? Sure. From a safety perspective it shows due diligence and efforts. But is it worth going out and buying a grad from a Ivy League school so you can check that box? I mean that's a signaling exercise.

[Dottie Schindlinger]

You know, it's. As you're speaking, I'm reflecting on how loaded that word expert is. Right. I feel like it gets so abused and misused. But I also do think a lot of boards, and we know this from our research, are eager to bring in quote unquote AI experts to help advise them. Do you have any advice for them on how to qualify who they're asking to come into the boardroom and what they should be asking them to help with?

[Josh Klein]

Yeah, yeah. And it's so I do quite a lot of this work and one of the things that I always, one of the things that I really appreciate is when I'm asked, what do I not know that I should. And then you can see they're evaluating my response to see if it makes sense. So there are a lot of self proclaimed AI experts out there that will come in and tell you that the singularity is nigh and you need to do X, Y or Z and it will save everything. Right. For one, I'm not so sure about the singularity and for two, there's no prescriptive answer with AI right now because we are all figuring it out. So I think that the real thing that directors need to do, particularly when they're looking for experts, is to see if they have an understanding of their business and can add some value. And that doesn't mean spouting a lot of, you know, acronyms.

[Dottie Schindlinger]

All right, so I'm always a big fan of the start, stop, continue model. Like, what should we start, what should we stop, what should we continue? So maybe give us a start, stop, continue. For corporate directors listening to this show as it relates to AI, start by

[Josh Klein]

getting a subscription to a model and asking it what they should know and have some conversations with it. One thing to stop is stop buying, how would I call it? Halfway models. So there's a lot of AI for X, right? Like here's a service that's AI for spreadsheets or AI for note taking or whatever it is. Like, figure out how to use the raw tools because it means you're exposed to a lot more of the rough edges where you can create more value. Stop, start and continue, continue asking questions like, you know, I'm really grateful to have an opportunity to have this conversation, but this is really where we're figuring it out. You know, companies don't know. There's no playbook really at this point. And it's the curiosity for listening to stuff like this and to ask questions and have conversations like this that are going to produce the answers.

[Dottie Schindlinger]

Well, Josh, before we let you go, there are three questions that we like to ask every guest who joins us on the show. You might remember these from the last time that you joined me on the but the first question is, what do you think will be the biggest difference between boardrooms today and 10 years from now.

[Josh Klein]

10 years. That is a big question. I think the biggest difference will be the number of humans in the room. Just to throw something spicy in there.

[Dottie Schindlinger]

Or do you mean the number of robots in the room? Yeah. That's great. What was the last thing that you read or watched or listened to that made you think about governance in a new light?

[Josh Klein]

There is an article recently by a gentleman named Yeg that was about the anthropic hive mind. And he was talking about how that company is run on vibes to a startling degree. And one of the things that came out of it for me was just how flat organizations can get and still produce large scale results. So from a governance perspective, and obviously this is anthropic. Right. They invented one of the leading models that's out there right now. But it, it made me realize how much the information transparency that AI can give us can create flatter organizations and different, more at speed organizations.

[Dottie Schindlinger]

And speaking of flat organizations, there's your own, right? I mean, I think you told me at one point that you keep thinking, okay, I need to hire some people, and then you just spin up another agent, realize you're good.

[Josh Klein]

Yeah, yeah. Although that has reached a limit, it turns out that I have to retain a number of subcontractors and some staff because at a certain point, human beings need to talk to other humans. You can't build trust otherwise.

[Dottie Schindlinger]

Yeah, that's a very good point. And finally, Josh, what is your current passion project?

[Josh Klein]

I'm working on a tool called Networker that analyzes all your past emails and communications and does a deep semantic analysis of who you know and how strong your relationship with is, is with them. So that if you're working on a current project, it can identify someone that you already have a trusted relationship with with expertise that you've forgotten about. So it's basically capturing that long tail of our networks because we all have massive networks and they're woefully underutilized. AI can help with that.

[Dottie Schindlinger]

Wow, Josh, make me a beta tester. I really want that tomorrow. That's awesome. Well, Josh, thank you so much for joining us on the show. It's been such a pleasure speaking with you. It's always great to pick your brain and find out what's happening. So thank you you so much for joining us.

[Josh Klein]

Thanks so much for having me. It's been a blast.

[Dottie Schindlinger]

We've been joined today by Josh Klein, board member, CEO and strategist with a proving track record of advising Countries and Fortune 500 companies on navigating disruptive technologies like AI. Josh, thanks for joining us on the show.

[Josh Klein]

Thanks again,

[Megan Day]

Dottie. I have to laugh about what you said earlier about having lots of conversations at the surface level. But then on the flip side, it's like I could sense your burnout in a lot of ways. This idea of going down this path of like, well, what is even an expert, Josh?

[Dottie Schindlinger]

Well, it's funny, right? It's so funny. You know, I think the first time I started asking that question was last year or two years ago, maybe. We had a speaker at our Elevate conference by the name of Sophia Velastige, who has been, you know, she was the chief AI office at Microsoft. And she really has just done a lot in the AI space. And she sort of suggested to me, you know, you kind of have to be a little suspect of someone calling themselves an AI expert. Unless they've, you know, worked at some of the big companies and really have been deep in the technology, I'm not sure they get to use that title. And ever since then, it's been something I've really wrestled with. Because the truth is, you know, when you're in a position of leadership and along comes a topic like AI that you know nothing about, your first instinct is, I need to get an expert in to advise us on this. And quite frankly, I think you kind of do the thing you've done with every other topic like that. You go back to the people you already trust and say, send us your fill in the blank expert to come and talk to our group. And, you know, it started to really bother me to think like, do they have any. Do they have any experts? Like, this might be an incredibly well respected firm. They might have done great things for that company in the past. But let's face it, I mean, Generative AI is really still very new. It's developing at a pace that we've just never seen any technology go ever before. Even Josh, who does nothing but this every day, said, I don't feel like I'm an expert. He's like, I was talking to someone the other day and they said, have you heard about this? And he said, no. And he's like, yeah, that's because it happened five days ago. And it was a really astute observation about just this whole conundrum around expertise. And what do we expect from the humans in this process? I don't know. It's something I'm going to continue to wrestle with.

[Megan Day]

I think, well, if the future is Vibes and Vibes only, then maybe Expertise Isn't needed.

[Dottie Schindlinger]

All right, Megan, I'm gonna. I'm gonna ask this because I'm an old person. Please explain vibe coding to me. I need to understand what this means.

[Megan Day]

It's just an unstructured way of prototyping is how I think about it, that you can build something that you going in and writing 7,000 lines of code. It's like, create for me this, this, and that, and it's based on vibes.

[Dottie Schindlinger]

I don't even know.

[Megan Day]

I don't know how vibes came about as the. The related term, but it. It tracks.

[Dottie Schindlinger]

It really. It really does. You know what? It makes a lot of intuitive sense. And isn't that kind of what generative AI is all about right now is intuitive sense?

[Megan Day]

There's a lot of detractors, like, a lot of folks that work in product management and product development and engineering, you know, have a lot of beef with the concept of, like, vibe coding and whatnot. But I think if you're operating at a certain level, if you want to test and iterate and explore new ideas at the pace in which the world now demands, I think it's pretty cool.

[Dottie Schindlinger]

Yeah, it is really cool. Well, okay, so I want to ask you your thoughts about. I mean, first of all, there was so much in this interview with Josh, but one of the things that he was talking about is just how it's so much more complex for governments, for companies to. To kind of govern AI because of the nature of it being a mimetic tool, memetic engine. And I really had him to, you know, kind of explain that to me and explain what does he mean by that. And it is really true that, you know, it basically behaves and reacts like a human being. And he brought up the example of. In cyber risk, right now that the majority of phishing attacks are now being launched by AI. The majority. I mean, that. I don't think I knew that. Did you know that information, Megan? That was news to me. I mean, I'm not surprised, really, but I did find that really interesting. And I think what's interesting there, too, is not only are the majority of phishing attacks being launched by AI, but now the majority of phishing response and cybersecurity response is being mounted by AI. So we basically have AI trying to scam AI. That's what's happening now. It's. It's wild. This is a wild time to be alive.

[Megan Day]

Vibes, you know, your. Your comment, and we have talked about this before, Dottie, around. How do people get the working experience to grow within an organization and elevate themselves to leadership positions. And you know, I, I think again, there's some really interesting points touched upon, but I do think, you know, some of the, the soft skills of leadership continue to be increasingly important and many would argue that that is not necessarily something fully taught by being in roles. It just sort of maybe comes a little bit more organically. So that just I think was interesting food for thought for me and the, the role of sort of inherent leadership in people as technology starts to embed more and more.

[Dottie Schindlinger]

Yeah, I really appreciated Josh's comment there that he really sees one of the biggest areas of enterprise risk is this what he called death by churn, where companies are kind of celebrating that headcount is going down. They're, you know, quietly not rehiring, they're using AI to fill in the things that a human used to do. But he's like, but at some point you're going to realize the people that you need the most are the ones that have been leaving the building. And that becomes a huge risk for companies. It's like, it's almost, it's almost like the corporate equivalent of population collapse. It's like eventually you get to a point where you just don't have enough human talent left to be a viable business. And it wasn't something I'd ever really thought about before. I mean, I think what we've talked about more in terms of talent risk is wholesale roles and wholesale professions being eliminated by AI. I don't think we've really thought about what this looks like at the individual company level where they downsize to the point where they can't thrive. And you know, it's so funny. Megan, have you watched the show Pluribus? I haven't yet.

[Megan Day]

It's on my list. But as you know, with a three year old, my TV watching is in like 20 minute increments once a day.

[Dottie Schindlinger]

Got it. Yeah. And Pluribus is not the show to watch with a three year old. But I've just say, okay, you haven't watched it and it's fine. What I will tell you is though, there's a lot of chatter online about whether or not Pluribus is kind of about AI.

[Megan Day]

Yeah, I have heard that.

[Dottie Schindlinger]

Yeah. And there's this sort of one piece of the puzzle where one of the people who has not been infected with the thing that has infected all the people on Pluribus is an author and she starts to work on her next novel and she's sharing notes with one of the hive mind people. And the hive mind person is so excited about there being new information. And it's kind of interesting when you think about that as a corollary for AI, right, that if AI knows the sum total of human knowledge, it's really kind of a database, a repository, a library of all the things that have been known up to this point. But that doesn't mean that it's necessarily going to be as creative or inspired or innovative or novel as a human brain would be, because humans put things together in ways that I just don't think to date, machines can do or have done, you know, and it's like, yes, you can use AI to create an amazing picture or video or a clip of audio. And it can sound so, so compelling. It can sound so much like something else that you already like, but it's not going to be completely groundbreaking or just totally out there. And I think that's the thing that we could end up missing. If our companies downsize too much and we don't have that, you know, that one wingnut idea that just comes in from nowhere and changes everything, we're going to miss out on a lot of innovation. So I think he's kind of right. That does pose a huge risk for companies, for sure.

[Megan Day]

I will say though, you know, when I'm inevitably out of a job, my backup plan, I think I'm going to be one of those bad bunny bushes from the super bowl and follow him around. That sounds like a pretty secure path.

[Dottie Schindlinger]

All right, Megan's going to be part of the bad bunny bush league. Meanwhile, I have just read about an island in Greece where they will provide free room and board if you promise to help them care for all the stray cats on the island. That is my. That is my, my post AI retirement plan. Going to go to Greece and help care for cats. All right, Megan. Well, this has been a little bit free range, but on that note, that wraps up another episode of the corporate director podcast, the Voice of Modern Governance. I'd like to say a few special thank yous, first and foremost to our very thought provoking AI expert, Josh Klein, Podcast producers Kira Ciccarelli, Laura Klein, Steve Clayton, our sponsors for the show, including kpmg, Wilson Sonsini and Meridian Compensation Partners, and most especially, thank you to Diligent for allowing us to have this free range conversation. If you like our show, please be sure to give us a rating on your podcast player of choice. You can also listen to our episodes and see more from the Diligent Institute. By going to diligent.com resources. Also, if you serve on a nonprofit or public sector board, then tune into our sister Diligent podcast, Leading with Purpose for expert conversations on governance, risk and compliance that makes an impact for mission driven organizations. Thank you so much for listening.

[Josh Klein]

You've been listening to the Corporate Director Podcast. To ensure that you never miss an episode, subscribe to the show in your favorite podcast player. If you'd like to learn more about corporate governance and tools to help directors do their job better, visit www.digent.com. thank you so much for listening. Until next time.