What lessons governments can learn from the EU AI Act

In this episode of the Better Innovation podcast, David Shrier taps into his expertise to examine the consequences of the EU AI Act.

Related topics

Join host Jeff Saviano and this week’s guest, futurist and innovation catalyst David Shrier, who draws on his expertise in data, fintech, digital identity, cybersecurity and innovation to examine the European Union's (EU) AI Act – a piece of legislation on which he contributed – and the precedent it sets for the U.S. and beyond. From delving into technology policy and development around the world, to artificial intelligence (AI) governance and ethics, Jeff and David present a comprehensive discussion spanning an array of timely topics.

To round out the episode, David offers a glimpse into his soon to be released book, Welcome to AI: A Human Guide to Artificial Intelligence, which touches on humans and AI’s unique capabilities, job security and industries poised for growth in an AI-driven future.

Key takeaways:

When formulating new regulations, it's crucial to engage a diverse array of impartial stakeholders, including think tanks, non-profit institutions, and civil advocacy groups, which operate independently from the commercial interests typical of the private sector.

There are certain professions, such as those in hospitality and healthcare, that will be more resistant to AI adoption because of the value placed on human-to-human connection within those industries.

The U.S. government should look to the EU as a model for how to utilize a principles-based and risk-based approach to regulating AI.

Transparency is crucial in interactions involving AI.

For your convenience, full text transcript of this podcast is also available.

Intro

Meet the people behind today’s leading innovations from the boardroom to the halls of government. Join Jeff Saviano a global innovation leader at EY. To hear from the trailblazers reshaping our world. You're listening to better innovation.

Jeff Saviano

Hey Better Innovation, Its Jeff. I had my friend David Shrier with us in the studio today, what a privilege of getting David here with us to do this live. He has got such a high degree of knowledge and experience in not just technology in general but specific to AI. David is a futurist. He is an innovator. He was in our market and Cambridge for years at MIT. Let me a look little bit more about David. He leads a venture studio today called ‘Visionary Future’ and he helps global 2000 companies create just breakthrough growth and incubate new technology enable businesses, but he is also a professor of practice at Imperial College in London, and through that appointment he has worked with European Parliament and their advisory committee for AI and he came on the show today to talk about policies as they relate to AI and also particularly to the EU AI Act. But this really only scratches the surface of David’s experience. He came on the show today to talk about his new book ‘Basic AI : A Human guide to artificial intelligence”. Here we go David Shrier.
David we are thrilled to have you on the show today.

David Shrier

Always a pleasure to be here, Jeff. You're looking tanned, fit and rested.

Saviano

I'll take any two of those three, but three. My gosh, that is that would be really nice. We've been trying to do this live for a while, and I have to say, I'm glad that we waited because we could have always done this via a virtual connection, but we didn't because you come through Boston and Cambridge a fair amount, so why not wait and do it live?
I think it's going to it's going to make for a better conversation.

Shrier

Absolutely. And, you know, the pace of change is ever accelerating. So, there's always something new to talk about.

Saviano

Always. And this is actually a line I'm glad you said that. This is a line, too. We have a special series. We're in season seven of Better Innovation, and we have a special series on AI. So, you're right in the middle of that. And you just wrote a book about AI. We're going to talk a lot about that. But you have been in this space for so long, it was really hard. I have to say. It was almost impossible to land on a single AI topic, so we broadened it out. We're going to call this the smorgasbord of AI outstanding.

Shrier

Well, I do like variety. You know, it's the family style meal. It's always much more fun.

Saviano

Yeah, I like that. As an Italian, I can appreciate the family style. And as we're in the U.S. here, as we're getting close, we're recording this just a matter of days before Thanksgiving. So maybe I've got food on my mind today. Okay, before we dive into this buffet of AI, let's start with the basics. Can you tell our listeners a little bit about your background? Let's go to David first and then we're going to get to some of the policy issues.

Shrier

Sure. Well, you know, on the air front, I wrote my first A.I. software program in 1991 when I was a student at Brown University, and it was something that I found interesting.
And then I set it to one side because back then AI was not terribly practical. It wasn't useful for very much. There were select narrow use cases that that that we could play around with.

Saviano

This is while you were in, in college.

Shrier

While I was in college.

Saviano

You did a major.

Shrier

A double major in theater and biology with a focus on neuroscience. So, I was interested in the brain and the technology behind the brain. And I was also interested in kind of people and storytelling and how these things sort of interact with each other. I actually wrote a play about Rosalind Franklin, who was the first scientist to decode the structure of DNA. And then, you know, I spent a lot of time in technology innovation. I spent now, well, we'll say politely, more than two decades of work trying to help management and CEOs, boards and companies to grapple with technology and how to implement and use it to grow revenue. And so that's been kind of a fascinating journey, worked on over $11 billion of growth initiatives, even briefly passed through the doors of EY, which was an educational experience to say the least.

Saviano

We love having alumni on the show. Everybody at some point will come through the halls of EY. So, we appreciate that about your background too.

Shrier

Yeah, no, it's a blue-chip institution with a lot of really bright people inside. And so that was, I think, a formative experience for me. But, but, but, you know, I remain fascinated with this question of sort of how can we use technology to solve big problems.
And so that's what I've dedicated my life to and through particularly through my work, first at MIT and then at Oxford and now Imperial College, which is which is part of my life, right, as a academic. So, spending some of my time in academia and some of the time in the real world, I've also started working with governments.
I've worked with over 100 governments on technology policy. And so today I help governments. I work, you know, in academic research. And then I also invest in companies and start companies. A lot of it is data driven, which kind of led me inevitably back to AI.

Saviano

I think that's what has always made your background so interesting to me.
You and I have known each other for you've also been a client, your client of EY, the CEO of a company that was a great client of ours. And I see as you have moved from academia and it's not as though you ever leave these positions, but you tend to take on a lot. So, it's not like you do these.

Shrier

The gig economy for the 1%, right?

Saviano

So it's like, you know, I got a bunch of things, but it's like you wrote the book about the gig economy, and I always appreciated that, that that I was thinking about that as we're preparing for today's interview of the many different hats that you've worn for stuff that we've tried to do together, whether it's a startup that you're launching about training in the metaverse, which was phenomenal, or your academic work was Sandy Pentland when we started our relationship with Sandy, I think that's about the time that you and I met and I just see as you've come in and out, not just the private sector, but from an investor, from an entrepreneur. And as a CEO, I think that that perspective, based on the breadth of AI, we were really excited to get you on the show to talk about that.

Shrier

Yeah, well, it's great to be here. So, let's dive in.

Saviano

All right. I want to talk I want to start the conversation. And given the focus of our talk today, could you give us a little bit of insight into your role and involvement with the European Parliament? You mentioned you've worked with 100, 100 different countries. You've spent quite a bit of time with the Parliament and more particularly the advisory committee on
AI. Talk a bit about at what juncture you worked with them and how you advise them.

Shrier

And so, I've been working with the EU in its different aspects now for, for several years. So, I worked on a project with the European Commission at one point around technology, innovation and growth and entrepreneurship. But what happened was several years ago now, it must have been three or four years ago I was approached by the Science and Technology Options Assessment Committee or Stoer and that was the EU Parliament Committee that most often dealt with advanced technology. And so, for example, they put together the blockchain legislation for the EU and several other important infrastructure related investment bills. But but they were the group that kind of took point on drafting the EU AI Act, which the European Parliament passed earlier this year and is now sort of being navigated through the different European governments and heading towards roll out in 2024. And so, they wanted to have expert opinions and perspectives from diverse inputs. So, so there were a couple of dozens of us from some pretty interesting backgrounds, not only limited to the EU. So, it also included folks from outside the EU like me, and they want to make sure that they were taking a thoughtful and expert informed perspective on how to legislate this emergent technology.
And it's important to note that we were putting this together before Generative AI became the kind of technology du jour. So, this was this was looking at everything that was exciting and powerful about AI up until sort of 2022, end of 2022. And I can share some funny stories about kind of what generative AI means for all that.
But just on its face of what we were looking at, we saw I was starting to have a bigger and bigger impact on society, and we saw that it was going to influence many different aspects of society. And so the EU very prudently has taken a principles-based and risk-based approach to regulating AI. And I think, I contributed to this perspective. But I'll say I emphasize that I think that's the right approach. You don't overregulate, you're not putting this massive book of rules in place, but you're saying, look, if there's risk of significant harm, like we're going to decide whether or not you can get a job by handing the decision to an AI, that's a significant risk of harm. Your livelihoods are at stake. That should be heavily regulated. Having a video game character that's controlled by AI that makes the video game more fun, no risks to harm. Yeah. And so, we shouldn't regulate that. And somewhere between the fun video game character and your ability to feed your family is at stake. Somewhere between those two extremes, there are varying degrees of regulation.

Saviano

And you agree with that. That risk-based approach. And if you look at just the way that you describe it, one example is if you're using AI to assist with hiring somebody, that's a that's a fairly high-risk activity because of the opportunity for introducing bias and discrimination. And so, you're right, a very and we'll talk about the U.S. approach and how that has been different. But do you agree with this risk-based approach?

Shrier

Yeah, no, I think it's the right way to go. You know, so if you're talking to an AI chatbot, it should disclose to you that it is a chatbot, but that's about it for regulation. It's like as long as I know I'm talking to a robot and it's answering my customer service question.

Saviano

No matter the risk. So, with that example, it could be high risk, could be, could be a customer service question that it's not going to, let's say, change the world if it goes wrong. But even in that situation and transparency is important.

Shrier

Yeah, transparency is important. Has to self-identified has to tell you I am a chatbot. You have to know you're not talking to a human being that's a moderate risk activity. If that chatbot turns into your interviewer for a job. Now we're getting into a high-risk activity and that's where it needs to be more, more tightly controlled. But I think what that demonstrates is, is a concept in policy regulation known as being, you know, prudent, not over controlling, not overregulating, but rather is having a proportionate response to the risk that is actually.

Saviano

Proportionality, I've heard some describe it as regulatory parsimony. Exactly right. Don't stifle innovation but regulate just enough.

Shrier

And yeah, and people forget that, by the way, people think of regulation as, this awful thing that's bearing down on me. One of the pillars, one of the four pillars of good regulation, is it should support innovation, it should increase competition. Those are, in fact, two of the four pillars we want to promote competition. We want to encourage, regulate innovation. And people forget that because they focus on the other pillars of regulation, which includes things like managing risk and promoting markets.

Saviano

They focus on those issues. And frankly, I think many people are blowing past the question of whether governments are going to regulate and to what extent they are. And we've been talking about it's, you know, part of the work that we're doing at EY in talking to a lot of boards and a lot of c-suites about deregulation and the suite of opportunities and risks and it's an incredible time that, that if you look at how AI is being used today and the risks that are there, you know, that is certainly a risk that boards have to consider for sure.

Shrier

And more than 100 governments are actively engaged in AI policy intervention. So, it's happening and it's happening, particularly in wealthy countries where businesses may have the largest percentage of financial exposure.

Saviano

And there's a lot of examples of governments that have very assertively said we're not going to regulate emerging technology. We've mentioned on one of our recent shows in the late nineties, President Clinton twice with e-commerce in 1997 and then with the Internet in 1998 with the Internet Tax Freedom Act decided to let industry self-regulate because the stock market was going through the roof at the time. But that was a conscious decision not to regulate. I think we are going to see many countries deciding to do something here, right?

Shrier

Yeah. And look, the Clinton magazine e-commerce principles which govern sort of, it basically is what enabled Amazon to become $1,000,000,000,000 company. I think it was a good thing we had this new technology, this new way of doing business. We didn't know how to handle it. And rather than put a massive pile of laws in place around it, we said, let's just make sure that we do the right thing here and we have a gentle or light hand on the tiller.

Saviano

Protect people, protect people in the process. China did it with the super apps and the government famously let them develop. First. We had Martin Shaw Zampa on the show who wrote this at the Peterson Institute, wrote this great book about the rise of the super apps. And we heard his story of the seven-year period where China said, we’re going to end it.

Shrier

There were a thousand flowers. BLOOM right.

Saviano

It's they bloom, but they also some people lost their life savings and it became at some point they had to step in.
But here and clearly within the EU, do you think that there's a risk of the EU overregulating with this approach, or do you feel like they got it right?

Shrier

Well, here's where the story kind of enters the second chapter, because so if you're going with principles-based regulation instead of rules-based regulation, you need regulators who are versed in the topic and able to be flexible and adaptive because they're applying their interpreting principles rather than just following a checklist. And in what's emerged with generative AI is you've got a probabilistic system, right? So, a simple way to think of Chat GPT for example, is it's a next word predictor. So, it looks at a string of words and then it says, Here's the probability that another word is going to appear in front of that string of words. And so that's how it constructs sentences and paragraphs and pages, and they randomize things a little bit.
So, what that means is every time you give it a prompt, it gives you a different answer. And the regulators don't like that because and this was a conversation one executive who's involved with a generative AI company was sharing with me was that basically he's sitting in Brussels, he's talking to the regulators and they're saying to him, but how come when we give it the same question, the large language model doesn't give us the same answer on the output side.
And so, he's trying to explain to them, well, it's a next word predictor, it's a probability engine. It's you know, it's not following a deterministic set of rules. It is following probabilities. And there, you know, the regulators, there's still some work to be done. So, one of the things I'm very concerned with is we have to build more capacity in these regulators. We have to educate them more about how to work with these systems and how they operate, because otherwise we're going to have principles-based regulation that can't contend with the pace of technology.

Saviano

They want to find that line. They want to find that line to promote competitiveness. They don't want the Eurozone to be a part of the globe where companies have a disincentive to invest because the rules are too stringent. But it sounds like you feel like they got that balance right by, as you said, mostly right.

Shrier

I mean, frankly, with what's happened with large language models in 2023, there's a bit of catch up to do and there's a lot of work to help the Eurocrats in Brussels get more conversant in these new technologies. Now, on the other hand, what's been happening, this is something I never would have predicted. I never would have predicted that Paris would become a major entrepreneurship and AI hub, not because, you know, Paris doesn't have great institutions or great research or whatever, but because the tax regime in France is so adverse to entrepreneurs. But yet, despite all that, Macron has been brilliant at attracting and supporting entrepreneurial innovation. And so, France and Germany have become major sort of epicenters for AI innovation.
And there are a number of companies that are growing up inside of these markets that are natively compliant with not only GDPR but the General Data Protection Regulation, but also with, you know, the expected sort of rules around European air approach.

Saviano

Economists perhaps don't love when the tax tail wagging the dog perhaps, but also, I think it highlights the complex role of government. It's not just the stick of regulation, but governments have used tax and other incentives to attract and to build industries. That's one of them. The questions that I love that we get in our lab at EY and Sandy gets and when you were at MIT, how many times did governments come in? And I can recall a meeting, we may have been there together, but Australian Government, there was a period where they came a lot and the question was.

Shrier

I set up that deal.

Saviano

You did? Interesting. Okay. So, you know exactly about and I always remember receiving that delegation a few times and they asked the question, how can we replicate Cambridge in Australia? We got that one of the members of the cabinet in Scotland came to us and asked that question.

Shrier

We get this question a lot. This was the conversation that I had with the European Commission over two years when as part of the project team and, you know, how do we fix Southern Europe, which economically has been a basket case for a while.

Saviano

And that role though, and so what you saw in Australia and what you're seeing in the European Union, it can't just be. But regulation has to also be about incentives.

Shrier

Absolutely.

Saviano

And consumption and government can lead by consuming this technology for the good of the people. Right.

Shrier

Yeah, well and what's interesting so I chose to move from Cambridge, Massachusetts to London to join Imperial College London in part because the UK has been a global leader in investing in AI, supporting AI entrepreneurship. So, a colleague of mine shared the statistics with me last year. Last year, 2022, there were about a thousand AI companies that got seed and around funding in Silicon Valley, and there were about a thousand AI companies that got seed and around funding in London. Now, that is staggering that you have an entrepreneurial innovation ecosystem that rivals Silicon Valley in such a critically strategic technology.
If you look at the stats on investment, the UK is the most productive nation on the planet per capita for AI innovation. So, if you look at raw numbers, the US is number one, China's number two, UK is number three or number four, depending on whose statistics you believe. But think about that. The population of the UK is only about 45 million people, and yet it rivals the AI productivity of much larger nations. And so. So, I moved over there to be part of that innovation ecosystem that AI revolution that's happening.

Saviano

And you think that the policymaker is or at least part of that equation,

Shrier

That's a good thing. No, I mean part of it. Why? Why is it there? Well, for decades since the 1940s, when the UK government funded the invention of A.I. technology in World War Two, UK has been committed to A.I. innovation.
And so, you have now universities like Imperial College, the number one research university in Europe, like Cambridge, Oxford, UCL, or a host of universities. Thousands of AI researchers across the UK, the Alan Turing Institute, Ada Lovelace. Their series of these government funded and government supported institutes that promote AI research and innovation that then sets the framework and lays the foundation for a host of entrepreneurs to then, you know, kind of get their startups going.

Saviano

Now, perhaps not so different from the approach to privacy and the approach to GDPR. And I know that that there was a session that Sandy Pentland led in Davos years ago that's been credited as starting that.

Shrier

Absolutely. It was 2006, 2007. I mean, he was early, right? I mean, he led the Privacy Working Group for West that led to GDPR. But, you know, the AI roots in the UK go well before that. So, it's an interesting market and I think it's a great example of how government can lead the discussion. And lo and behold, more recently, earlier in November of 23, the UK government led the first International AI Safety Summit where 30 countries came together along with private sector, and started having a serious conversation about what I called, you know, humanity's greatest existential crisis, which is this technology that on the one hand could completely destroy the society we live in, and on the other hand, it could solve some of the largest problems that we're grappling with.

Saviano

I just saw this statistic that sometime in 2024, 75% of the world's population will live in a country that's governed by modern data privacy rights. That is GDPR, but it's also where it's been extended to Canada and California and Brazil and UAE, Australia, etc. And so, some are starting to talk about the EU AI Act having the similar trajectory.

Shrier

Kind of a lighthouse effect.

Saviano

Yeah, it could, yeah. What do you think? Do you think that that's possible? It's certainly possible.

Shrier

I mean, we're seeing serious AI, so the big economies in Europe, France, Germany and Italy have come together and agreed on some more specific rules and principles around how to govern, manage AI, how to, how and when to intervene. So not getting too involved too early if there's a problem but having a clear sort of pathway to addressing problems. And that in turn is going to set the tone for the EU and then the EU, I think will certainly set the tone for a number of countries. But, you know, there are other countries that are forging their own path.
And so, on the one hand, China has been doing a lot and in fact they've been doing, China has arguably the most progressive regulation as long as you don't care if the government knows everything about you. China has done among the most to protect its citizens from private company depredations. Another path is being forged by Malaysia. So, we just kicked off a new initiative, Imperial College, called the Trusted AI Alliance and we're working with the government of Malaysia through their national payment network, through payment. And they are they are creating, call it a non-aligned nations approach to addressing technology disruption like AI. And so, you know, they're not going to just copy and paste the EU AI Act. They're going to come up with something that makes sense for the Malaysian context.

Saviano

And the US won't do that either. We'll talk a bit about the U.S., but just but just to stick with that, that region of the world is important work happening in the Philippines enacted for laws in 2022. Japan also second largest IT Sector of the OECD countries, huge and making big r&d investments but also
,

Shrier

Even elected an AI mayor in one of their cities. I didn't see that. So yeah. Well, I think it was a more of a ceremonial position but, but nonetheless and I ran for office like an avatar.

Saviano

Yeah. That kind of thing. Yeah. Okay. Well, that could promote some targets. Certainly, raise awareness.

Shrier

Japan is a country that is very comfortable with technology, and the regulators have put a lot of time and thought into it, especially the question of allowing copyrighted material to train AI systems. And they were way out in front on that issue. Many believe that that's the right, that's the right approach and probably the right approach in the U.S. Let's stick with the U.S. And because the approach has been so different and compared to the EU.

Shrier

I would use the word dysfunctional but go on.

Saviano

Well, well, I mean, I have a lot of opinions. Explain what you mean. Explain how you feel the approach has been in the U.S. Now, as we're recording this today, it's only been a few weeks since they issued the 100 plus page executive order that came after the AI Bill of Rights, the voluntary agreements with big tech.
How do you view US policymaking right now with AI?

Shrier

Yeah, so there are a couple of problems that are cropping up already. First, let's look at the numbers. Right? I was I think what was it? Mark Twain said something like figures don't lie, but liars’ figure. So, you know, the budget request for 2024 the Defense Department ask is $842 billion US and the allocation for AI R&D is 5.5 billion. So that does not include select secret military programs around AI. And I don't think it should. But, you know, there is a little extra money that that's kind of under wraps, but that by and large, you know, we're spending far more on defense than we are on AI R&D. And I think that's a problem. And, you know, I say we although I live in London, I'm still a U.S. taxpayer, and so it's still my money that's going to not underwrite enough AI R&D. The second thing that's happened is so the executive order and some of the conversations in Congress have happened in a less consultative fashion than we recommend as best practice. So, in theory, when you're adopting new legislation, you should have town hall opportunities where average citizens can comment. You should have multiple rounds of industry engagement where different participants in industry can kind of you have small players, mid-sized players, large players, and you aggregate a significant and diverse body of input in putting together new regulation. And what's been happening is, I'm not going to go all the way to saying that there's been outright regulatory capture, but we're getting very close to that where a select handful of very powerful interests are having a disproportionate influence on what's getting adopted. So, there were a series of consultations where only a very small number of very large AI companies were invited to help then influence what rules and recommendations were put together. I think that's a problem. There was not enough involvement from neutral parties like think tanks and non-profit institutions, civil advocacy groups and academic institutions who don't have the same sort of commercial interest that the private sector does. The second thing that's happening, which is a problem, is there's a there's a think tank. And so, I won't name names that I have, you know, given the nature of this forum, but there's a prominently funded think tank out of California that has been underwriting a series of AI policy interns basically have been paying for air experts that they train and fund to go work for free in the offices of a number of prominent congresspeople and senators, as well as I think also the federal government. I'm not sure on that last part, but you essentially have a private lobbying effort that isn't registered as a lobby that is going on, that's influencing how the legislation is getting written and adopted. That is a big problem. It's not transparent. It's being underwritten by some very wealthy individuals. And, you know, that results in legislation that benefits the few and not the many. And so that's also a significant problem.

Saviano

Perhaps not as different as other technologies or maybe it's more pronounced here. I can.

Shrier

It's very pronounced here.

Saviano

I can give an example of what happened with privacy regulations where privacy enhancing technologies, pets became an opportunity to comply with complex privacy regulations that may have benefited some tech companies. Is that an example? And do you think that what's happening now is materially different from what we saw with privacy?

Shrier

Well, you know, you just you have a non-transparent agenda that is getting embedded into legislation. And so, I do think it's different and I think there's more money at stake. I think there's more influence on people's everyday lives at stake. And there's systemic risk to the economy at stake. And so, yeah, I do I do think it's different and I do think it's a problem.

Saviano

We're still at a pre legislative period in the U.S., even with 100 plus page executive order that calls for a lot of studies and it calls for an analysis of how the US government can consume. What's your point of view on the scope of the EO?

Shrier

Yeah, So, it's good that the U.S. has done something, you know, I mean, there's a real problem when the U.S. didn't do anything. You can't wait. You have to do something. You don't want to regulate in haste, but you've got to do something. And so, I think it's positive that the U.S. is doing something, but then if you peel back a layer and look at the substance, for example, the specifics of how we, so we have a consolidation right now where 5 to 8 companies essentially control the foundation models that everything else runs on.
And right now, it's been left up to a closed-door negotiation between the Commerce Department and the 5 to 8 companies as to how commerce is going to regulate those companies. And again, this lack of transparency in consultation, I think it is concerning. It's not that you shouldn't talk to them. Absolutely. They need to be part of the dialog, but other people need to be part of the dialog also.

Saviano

One of the eight core themes of the executive order is to promote innovation, to promote competition, including small, medium sized businesses. That was a step in the right direction.

Shrier

Absolutely. I think that's a positive measure,

Saviano

But perhaps didn't go far enough.

Shrier

Yeah, and you know, there's the statement and then there's what reality is backing it up. And right now, Washington is in the middle of a perpetual it's turning into an annual dysfunctional crisis around the budget. So how much money is actually going to go against that competition initiative at the end of this very contentious budgeting process? I don't know.

Saviano

Is it more of a positive step in the right direction, this EO or do you feel like there was some missed opportunities?
Where are you on that spectrum David.

Shrier

I think it's a first step and so I'm glad that some step was taken. I think there's a lot more work to be done.

Saviano

The voluntary agreements didn't do much for you with big tech.

Shrier

Not really. I mean, we've already seen what happens with social media, with big tech when it's left to its own devices. And so unfortunately, I'm a free market capitalist. Like I believe in the benefits of funding innovation and having innovators go do things. It's not that I'm begging that we have a lot of government intervention, but at the same time A, we saw with privacy and personal data and social media, what happens if you just leave it entirely to the industry to sort itself out and B I direct you to Mariana Mazzucato is excellent book, The Entrepreneurial State, where she talks about how the role that government has played in funding a lot of the innovation that we rely on today. And so, you know, I think there's there is a role for government. I think that it needs to be a little more rather than less on AI because the dangers are so great and the opportunity for humanity is so great.

Saviano

Before we leave this discussion of policymaking. And then we've been at least some focus on the EU. Yeah, in the U.S. Why do you think the EU approach is so vastly different from the US?

Shrier

For one thing, you know, we've had for several decades now an active campaign in the US to essentially destroy the organs of large government. And there were some rationales for it. Suddenly, the idea that government is bad emerged like, we should get government out of our lives, destroy government, get rid of big government, everything should be, you know, libertarian and for the individual. And there's a lot to be said about that. And probably a longer conversation as to why that's based on some fundamental economically flawed fallacious arguments. But in Europe, people recognize government as something good that delivers good things for your life that you know, you have, for example, universal health care, where people don't have to go into debt if they get sick and everyone knows that, you know, they might have to wait to see a doctor. But everyone knows they can see a doctor and they don't have to worry about, you know, mortgaging their house if they have a serious illness. And that's just one example of many where people actually think of government as a positive force in their lives. And so, there's less resistance to the idea that government should have a role in things like what technology means. And AI in particular.

Saviano

Just last season we had as one of our guests on the show, David, Elizabeth Grenier, as you may know, Elizabeth, she was she was here still around here, but now affiliated with Oxford. She's a senior research associate at the Institute for Ethics and AI at Oxford. And she wrote this great book called Beyond Data Reclaiming Human Rights at the dawn of the Metaverse. She said, right where you are, which she was on the show, and she talked a lot about that, how World War II is not such a distant memory, and you compare it to privacy legislation could point to that period of time as still significantly influencing societies in Europe and what their expectations may be. Perhaps that also will influence their views of AI, too.

Shrier

I think that's right. I mean, it is important to recognize that even within Europe there are different views on privacy. So, Germany would rather sacrifice some personal safety in the name of better privacy protections. London, certainly. But the UK has more invasion by the government in your privacy in the form of all these CCTV cameras in the name of more safety. So, when the July 7th bus bombings occurred, they were able to track down the bombers in three days because they had all these CCTV cameras. So, there are differences of view in Europe, but on average, the Europeans tend to view government as a source of something positive in their lives, much more so than we do in the US.

Saviano

I think that's a good segue way. David. Let's pivot to your latest book called Basic AI A Human Guide to Artificial Intelligence. It comes out just a few weeks, a couple of months, right.

Shrier

Just so it'll be out January 11th in Europe. In the U.S., it's being released as welcome to AI, a Human Guide to Artificial Intelligence. okay. From Harvard Business Publishing. But my publisher in Europe is Little Brown. And yeah so, it's the same book, just different title, different cover. And either way, the goal is to try and help the average person, the average businessperson, average individual understand, you know what can, what is AI? What can it do for you?
What are the risks to your career? And how can you position yourself to succeed in the new AI era? And then where are there opportunities for the betterment of humanity? And what role should government play in that? So, it tries to cover a lot in in a very kind of compact package. It's almost like the fly by version. So, every chapter begins and ends with a little executive summary. So, if you just want to skim the book, there are these little bullet point summaries that help you get the highlights. But it should be enough to get you started and give you equip you to think a little differently about A.I.

Saviano

You did cover a lot in the book, and first of all, our team.
thoroughly enjoyed it. Thank you for the advanced copies. It's hard to sign a digital PDF, but when the real book comes out, you'll promise to give us perhaps a couple of signed versions of the book.

Shrier

My pleasure.

Saviano

I look forward to that. The team loves it. It's just an excellent guide for people looking to get started. And we talk about boards and senior leaders and in our travels, people don't know where to start. They don't know the basic or the welcome guide to how do I even get started? There's so much in the book. I don't want to give away all of the tips and tricks. So, let's focus the remainder of our time, if it's okay. David, on just a couple of topics and the one that you just hit on, I want to dive into a little bit the impact on jobs.
Yeah, and impact on the work. Profound, by the way. Yeah, profound it's probably the most significant question that I'm getting from boards is about responsibility to the employee base, which I think is a good thing. Also, how do we retrain and how do we retool? I just want to quote from the book, if that's okay. You write in the book, quote, “Prior to generative AI experts were forecasting up to 50% jobs displacement within the next five years, disproportionately impacting the global South. New analysis reflecting generative AI shows 80% jobs displacement with more job categories in developed countries under threat.” end quote. David. That 80% displacement could be catastrophic on the labor market. Talk about that degree of impact.

Shrier

So, we have to almost reason by analogy because we haven't really conceptualized that kind of labor disruptions since World War Two. You know, we've lived in this expansionary economy, an economy in of abundance, frankly, for a long time. And so now, so first of all, it's worth mentioning that jobs making more than $100,000 per year US are among most under threat from air disruption.
Now that we're in this generative AI world. And so, tax accountants, management consultants, economists, lawyers, people who previously were immune to AI disruption, the white-collar jobs are now the ones squarely in the crosshairs.

Saviano

I'm getting squirmy just sitting here listening to you. I felt like the tax and accounting, it's those are areas, of course, -very rules-based systems.

Shrier

Absolutely

Saviano

right. And that's the first place.

Shrier

So, EY has how many what, 600,000 employees?

Saviano

We're well under 400, but both rendered an 80,000 plus. Well, we have about 70,000 tax professionals.

Shrier

All right. So, 380,000 employees. That is why probably the right size for the firm is 40 to 50000 people. And a bunch of EY for a variety of institutional reasons, is probably never going to go there in the next year or two.
But you could. And if you don't, are your competitors going to and I know for a fact certain that some of your competitors are heading exactly in that direction as aggressively as they can. And so, the resistance, of course, as is the nature of these institutions. So, you know, the question is, what do you do if one company fires even half their employees, that company might see a huge increase in in earnings if AI successfully stepped into the gap and it fills those jobs. But if every company does that, you destroy the economy. Now, you don't have people earning salaries. They then are not consuming goods and services and there's no economy.

Saviano

This idea of and we can talk a bit about the ethics of AI and what some of the issues are, especially for the private sector. You're a board member and you have a fiduciary obligation to your shareholders, but also an obligation to the world. And, you know, let's face it that these machines are working for us at least today anyway, friends. And if we want to advance the notion of human flourishing, hopefully then we're going to see it continue.

Shrier

I mean, you have this interesting question between shareholder capitalism and stakeholder capitalism.

Saviano

Yes.

Shrier

And You know, The CEO of a major company in the UK went in front of the press and said, I'm going to fire 42% of my workforce over the next seven years. I'm going to replace them with AI. And so, he is going to get some wonderful margin expansion and stock price expansion for increasing the profitability of his company. But it's going to completely ruin the entire country if everyone starts copying him.

Saviano

And that was a bold move. And we've seen other CEOs, perhaps a little less bold, say things like we're going to stop hiring in one particular division, we think.

Shrier

Or across the company, in the case of one large tech company or across the company.

Saviano

And you're right, and I know who that is. And there are large companies that are either they're freezing hiring or they're, we hope, proactively looking at what are the skills and capabilities and how can we out and we're going to come to this because I think it's important. Before we get to some practical tips for people. Sure. Ensure that you're not the ones who are disrupted. You explore a few industry examples in the book. There was one I thought was fascinating around investment banking. Talk about how you think, Let's just take that profession. How do you feel AI will impact investment banking in the next, let's say, 3 to 5 years.

Shrier

In fact, is already impacting it. And it's really creating some challenges. So, the old model of investment banking was shaped like a pyramid, the labor force shaped like a pyramid.
So, there were a lot of analysts, the base of the pyramid and a reasonable number of associates above them, and then a smaller number of vice presidents and then above them, perhaps at some banks directors and then above them a smaller number of managing directors. And maybe if you know, particularly in the last 20 years and even smaller number of senior managing directors. And so, you have this up or out philosophy of you had to work for a couple of years as an analyst and you either got promoted or you left and the pyramid sort of trimmed itself, but it still operated as a pyramid. And what did those analysts do? Well, they would work on financial models, they would work on PowerPoint decks, excuse me, they would work on financial models. They would work on slide presentation.

Saviano

They paid their dues.

Shrier

They paid their dues. They learned the culture of the firm. They learned how client work happened. And so, they kind of became represented because if you think about this, the difference in terms of capabilities between any one of the top three or five investment banks is indistinguishable. The way they differentiate with clients is on their culture and how they present themselves and how they do their work and the relationships that they built.

Saviano

You referred to this class as the elevator asset businesses.

Shrier

Absolutely, because every day your assets walk out through the elevator and leave the building. And so, if you replace all those analysts who are doing the slide presentations and the spreadsheets and things with AI, which a lot of people are doing because it's faster, it's cheaper, it's more consistent, its higher quality, then who do you promote to vice president? And so, their actual they rapidly these investment banks rapidly adopted technology. Now they're starting to face bottlenecks around who can we promote from within that has been trained in our culture if we've shrunken the size of the pool.

Saviano

To a lesser extent, they face those issues ten years plus ago when they offshored some of their work, right. When they had they went from that for a U.S. based investment bank and they pushed I don't know, but let's say was 30% of their work offshore. They would they also faced some of those succession planning issues, but not like this.

Shrier

Now, it's much worse because the scale, the magnitude of the change is much greater. Yeah. And so, they have to rethink that. So, if you were to design an investment bank or a management consulting firm today, knowing what we know about generative AI, it probably would look like a very small, inverted pyramid, would probably have a handful of very senior people who would have relationship with clients because a lot of big business is still sold on relationships with boards and CEOs.
You'd have some mid-level process people to basically manage the eyes. I think you have a bunch of I that that business would be dramatically higher margin than the management consulting firm the investment of today. But it does have a succession problem of okay but who do you then promote down the road? Who do you promote.

Saviano

And we thought and it's so interesting that you phrased it that way because I think conventional wisdom said that, you know, when firms like that were disrupted, you would move from a pyramid to a diamond. And I think what we're finding is that maybe it's not so much even a diamond with the bulk in the middle, that what Generative AI do is kind of go right after those knowledge workers in the Middle East. So now you're not going from a period from a pyramid to a diamond. You're going from a pyramid to an inverted pyramid.

Shrier

That's right. You think once you're 100% correct, that's what I'm seeing happening. By the way, it's roiling the industry. And so, your competitors, certainly I can speak to, have been going through executive turnover in large numbers as there have been pitched battles in the c-suites over how to deal with this transition.

Saviano

To some extent, law firms have been dealing with this and we've been spending a lot of time with legal tech providers. And yes, and the class that I teach at BYU and the law school just last week, we had six vendors, legal tech vendors come in. David. Every one of them shifted their business model in the last six months to significantly include generally that's not surprising.

Shrier

But the main thing, the main barrier with the legal profession is going to be lawyers. They're going to they're going to follow the rules, right? Yeah, but they're going to fight. The rules change that would allow the technology to be used. But, you know, you can put guardrails in place to stop the or, you know, manage the lens, hallucinating, you know, making up things and if you do that, you know, legal tech can replace most of the lawyers.

Saviano

There's some aspects of that I loved. We had an e-discovery solution and it's imagine a complex lawsuit that could be 10,000 plus documents that you have to pass and you have to figure out. Yeah. And impossible. I remember the early days of my career being thrown in a conference room with, you know, 30 M&A agreements saying find the same six clauses in these 30. We just don't do those things anymore. And there was it was all computers. It's and it's better it's a better in this age of collective intelligence that's the best but still and this is what we promote in the class and with my students, it doesn't mean that lawyers are going away. The empathy and the judgment and what it means to be a human lawyer will remain but significantly impacted by this technology, don't you think?

Shrier

Absolutely. Well, it's a look. The good news, such as it is, is that there is certain things AI is terrible at, like empathy, like teamwork, like creativity. And those are examples of some of the kinds of skills and capabilities that people need to lean into in the face of AI disruption. As far as the legal profession concerned, I mean, that that's really going to be a battle over time as some law firms begin to adopt more AI and then kind of push the legal profession as a whole to adopt more AI. But it's also different overseas. In the UK, the law firms have been able to be more aggressive and it's interesting to see that play out.

Saviano

If we could, let's stick with these issues for workers. And yeah, I have to say that of all the chapters, chapter four is my favorite chapter in the book, and that's it's called Reskilling and Developing Cognitive Flexibility.

Shrier

Absolutely.

Saviano

What do you mean by cognitive flexibility and how is it relevant to this AI influenced labor market? David, talk about that.

Shrier

There's a concept known as artificial general intelligence or AGI, and it's the holy grail of artificial intelligence research in computer science. And the idea is you have a computer that can think like a human being and think it can be faced with any general-purpose task and it could eventually figure it out. That's AGI. Now there are people in Silicon Valley who are throwing the term around as, we've discovered AGI. They haven't discovered AGI; they're still working on it. We're still a ways away from it. My colleagues at Oxford would say that, like Stephen Roberts and Oxford would say we're probably 200 or 300 years away from AGI. My contention is we're going to have something that looks a lot like it, like we wouldn't be able to tell the difference in more like ten or 20 years. And so that's a that's a sort of like he's correct. And yet there's this other thing of the functional AGI. But in the meantime, certainly for the next couple of decades, there's certain things that I just can't get right. It just doesn't do very well. And people do very well. And so, people are great general purpose thinking machines and people have the ability to engage in tasks that require empathy, that require the acquisition of new information that AI is just much less good at. And so that's where, you know, I think that we have an opportunity. And one of the things I push is kind of flexibility is a fancy way of saying stretch your brain so that you acquire the skills that you had as a young child of learning new things quickly.

Saviano

And I have to say, I, I read I read chapter four more as a parent than I did as a professional.

Shrier

Everyone's boy ever asked me like, what should my kid study in college? Yeah, well, that's I think that's the chapter.

Saviano

And I also I have I've got three kids. One is an actuary. So, we've had some really interesting discussions about that. That is like a job under threat. Yeah, it is.

Shrier

Absolutely. It is a job under threat.

Saviano

But this notion I want to stick with this notion of cognitive flexibility and you provide some great tips in the book. And I think this is applicable not I mean, this is not David projecting 200 years out to your last AGI comment about general intelligence. It's about like what's here today.
If you're a 20 something like my kids are and you're looking at it, you've got the better part of 30 years in your career. What can you be doing in those tips about being flexible? And you just said a few of the tips about always learning, right? Constantly learning. But there was an element that I love. Maybe it's the tennis player in me, but you didn't think about. But maybe you did think about sports when you were writing this chapter about this notion of practice. And there's this concept called deliberate practice and the importance of practice. I feel like and I, I have been working on my stupid backhand volley, like literally for 30 years. But and so from a sports perspective, in music, like we get what practice does, but something happens to us as professionals and like we turn off this notion of practice and I was so interested in your use of that word as a tip.
Am I being I overreading too much into it or is practicing this, this learning and flexibility an important part of the equation?

Shrier

The best way to learn is to apply knowledge, right? You don't want to just memorize facts. That's not learning. But if you are trying to work through a new concept or acquire a new skill, trying it out again and again,

Saviano

The muscle memory.

Shrier

The muscle memory is incredibly important for acquiring that new skill. And so, that's the discipline of practice.

Saviano

You have to actually engage in it on a regular basis. If you do it once and then you come back to it three months later, you're not getting the cumulative effect of getting better at it over time because you're building these problem-solving skills. Right? And that's a big part of it.
You describe in the book of Cognitive Flexibility is that and it could be, you know, doing brain teasers every day could be something that's really good for you, right? Absolutely. But you've got a practice, this ability to solve problems in new ways, because those are the skills that will survive, that I will never be able to touch is, as you said, empathy and judgment and the ability to work in collaboration, teamwork, collaboration, problem solve.

Shrier

But learning how to work together as a team is actually another skill that you have to practice. If you've never worked as a team before, then you know when you're thrown into your first team who leads, who follows? What's your responsibility? How do you learn to trust and rely on your teammates to get their piece of work done? There's a whole host of things that crop up if you haven't developed that teamwork muscle. So, and it's incredibly important because of that concept of growth that I talked about, that new technologies are coming along, getting adopted with great rapidity. And therefore, what does that mean? That means that you're going to have to learn a new skill every 3 to 6 months. And so, if you're in this continuous learning mode and you've got an active set of mental muscles where you've learned how to learn again and you've been practicing, and so you've gotten into that habit of knowledge acquisition and you've been working with other people to engage in peer learning so you and someone else are sort of talking about and think through the idea together you're going to be better equipped to deal with the next 20 years.

Saviano

But sometimes the hype of emerging technology gets way out of control. If it just so happened. We're recording this November 2023. If we were recording this 18 months ago, you and I could have been having a conversation about the metaverse. We could have, right? This feels different. It's almost certainly different, right? How would you explain it? But it's easy now to look through the rearview mirror.
Sure. Okay. And I know that you're also involved as an investor, as an entrepreneur with some metaverse activities. I think many of us believe that would have gone faster earlier. But if we were sitting here 18 months ago, we would have been more hyped on it than we are today for sure. Right.

Shrier

One thing that's important to recognize is the difference between what's actually happening in the world and what the media thinks is interesting.

Saviano

That's a good place to start.

Shrier

And so, the metaverse as a as a market sector category is $66 billion in revenue last year in 2022. So, people reports of its demise are greatly exaggerated because it's no longer an exciting topic to talk about. The media has been busy trashing it, but you know, some companies are trying to rebranded as facial computing.
The key thing is you've still got tens of billions a year of revenue going on in the metaverse. The thing that's been very slow to get adopted is virtual reality. Yes, but that is only one tiny segment of this overall market.

Saviano

Well, Web3 has been advancing in certain places.

Shrier

Web three has been advancing and augmented reality is advancing. And I have a portfolio company that is they signed a contract with a major casino chain to do a show in Las Vegas. And so, they're going to have a lot on in 2024. In 2025, they're going to do, you know, eight figures of Ebola off of Metaverse based on this contract that they've signed, events event driven. In that instance, its mixed reality, event driven. But on top of that, they have another eight-figure vibe off of their augmented reality game that you'll be able to play for your phone.

Saviano

There's just a way for you to use that theater degree from Brown Perhaps, perhaps this is where we're going to see. David.

Shrier

Yeah, I guess I've come for I actually, oddly enough, I did in fact produce my first produce. I think I got a producer credit, my first independent feature film this year.

Saviano

Did you really?

Shrier

Yeah, which is a very small investment. But a friend of mine was putting the movie together and it's got some brand name stars. But I guess we can't name the film.

Saviano

I can't wait to see that after the show you can.

Shrier

Sure.

Saviano

Okay, I want to round out this discussion. And so, for those early in their careers and again, they're looking at this as a parent wondering what the future work might look like for their kids, what industries or jobs or capabilities like. I understand the need to be cognitively flexible, but what about like real jobs and capabilities? How would you advise somebody who has 25 to 30 years left in their career?

Shrier

Yeah, well, and I have I have teenage children so I think about this myself. So, on the one hand, you know, the old joke to a man with a hammer, everything looks like a nail. Yes. And so, since I teach entrepreneurs, chip at a leading research institution, you know, I always think that being able to create your own wealth is a fantastic skill to have. And becoming an entrepreneur is a great thing. So that said, there are certain professions that are going to be more resistant to disruption from artificial intelligence. Hospitality is one of them. So, people still like going to restaurants and having human servers and interacting with people. So, at the absolute bottom of the pyramid, you know, the fast-food restaurants have been adopting these order screens to get rid of cashiers. I'm not talking that part of the market. I'm talking kind of fast casual up to high-end, high-end hotels, etc., with personal connection matters. So that's one health care in particular is another one where, again, people want a human doctor or a human nurse. They don't want to have a robot, you know, that that sort of treats them. So, those are a couple of instances, a manual labor profession. So, things where you're actually physical construction is a long way away from effective robots. So those are three professions where you've got at least a 20-year trajectory.

Saviano

You didn't say tax law.

Shrier

Tax law should have been automated ten years ago, but that's another story.

Saviano

That may be another show. We can have a good discussion about that. David, as we're getting close to the end of the show, I want to make sure we talk about the importance of governance and ethics. And I'm curious about another one of your ventures, the Trusted AI Alliance. Talk about the Alliance and how it may help with AI governance.

Shrier

Absolutely. So trusted AI Alliances is being led by Imperial College London. We have several other major academic institutions who have also joined us in this effort. So, more than 3500 AI researchers around the world who are working on problem of building a better future through safe and responsible AI. And so, the three areas we're focused on our air systems assurance, which is a fancy way of saying, you know, making sure AI does what it's supposed to and not what it's not supposed to. Workforce of the Future, which is this question we were discussing of what do you do if you disrupt 80% of jobs and how do you reskill people and what sort of social safety net you put in place around that? And then the third is AI for Humanity, where I can actually help address critical problems of the problems of climate or human longevity or what have you. So, within this, you need better policy and better governance. So, corporations, on the one hand, need to understand how do we better manage A.I. disruption? What are the rules procedures, policies we can put in place so that we are both compliant with regulation like the EU AI Act on the one hand, and also that we manage the risks and capitalize on the opportunities on the other hand. And then the second part is obviously the government policy side. You don't need to reinvent the wheel. There is a lot of prior art on technology policy regulation to leverage. And so how do we put that to work against the AI problems?

Saviano

But this is community driven. And you mentioned, I forget if you just mentioned it here or in some of the stuff, we're reading to get ready for today, 3000 members of this community coming together that and to go back to the points you were raising about the U.S., that's kind of what you were hoping and talking about the U.S. would do, right?

Shrier

Yeah. Well, by the way, that's just the academic members. Then we also bring together non-profit and government and civil society, NGO and industry and the big companies that are making the tech. So, we bring everybody together around the problem.

Saviano

I loved one of the projects I was reading about The Guardian on A.I. project, which at a high level you could have one A.I. system that sort of a watchdog for another AI system. It can be an early signaling device.

Shrier

Yeah. So, remember the old saying, who will guard the guards? Right. And the answer is you're going to need A.I. to monitor and help manage AI. And there are a few ideas we have about how to actually put that into practice.

Saviano

Are you searching for other members of this community to come together? Is there a message to our audience, if they're interested, to get involved?

Shrier

Absolutely. So, if you if you are a large company or medium sized company or you are a government or you're an innovator, we'd love to engage with you and have you join us in building this better future for humanity.

Saviano

And ethics, of course, is going to be critically important. We've mentioned a few times and some of the work that we're doing, talking to lots of boards, lots of members of the C-suite. I'm reminded that the average age of a corporate director is in their mid-sixties ish. It's not exactly the technology adopting millennial generation, but they have seen a lot of technology over the last 40 years play out in society. So, there are opportunities to reason by analogy. So, I am doing a lot an increasing amount of board education to, you know, a concise one- or two-hour session, give them enough comfort, confidence and capability so they can ask the right questions of management to get intelligent answers out.

Saviano

As you have had those sessions, what do you come away feeling about boards and what's missing today in governance and ethics as well?

Shrier

Yeah, I mean, they're all looking for where to get started. They're all looking for direction. What do I do? First, they know it's important and that's good. Awareness quotient is high, so they know it's important. Then the question becomes, what do we do about it?

Saviano

And that's where, you know, they have some gaps as to where to get started. And that's the opportunity to engage in discussion and capacity building. I think we're going to find that that for most companies, I will intersect with what the courts call mission critical activities. And if it's mission critical, then there's been a series of cases since a landmark decision in the mid-nineties involving Caremark. Those cases say for mission critical activities, boards have to do more.

Shrier

There should be a board subcommittee. There should be a chief officer who reports to the CEO and shows up at board meetings periodically. There should be clear rules and governance. There should be third party audits that are conducted of the algorithms and activities on a periodic, if not continuous basis. there's a lot that can be done which doesn't require reinventing the wheel. This is this is off the shelf instruction that boards can take.

Saviano

We just actually spent some time over the summer as a firm. We studied the S&P 500 and there are 67 companies of the S&P 500 that have a tech committee. That percentage is actually much higher for the Fortune 100. And what's also interesting, the I think it was the National Association of Corporate Directors did some research and they found that companies that actually have a tech committee perform better. That the financial performance against the New York Stock Exchange composite or NASDAQ is better. Not surprising, right? You put the attention on technology, not surprising that you'll see it on the panel's side as well.

Shrier

It's irresponsible not to.

Saviano

And that's just and we can think about it from not to oversimplify, but there's a legal compliance layer. What do the courts what does the law say? You as a board need to do? But then as we're hearing, I've heard from so many boards and so many business leaders that they want to do more, they want to act responsibly. Even a simple example, the research shows if you want to discover bias in a model, the more diverse team you have studying it matters. You and I have talked about that for innovation. I mean, that's we've known that forever for innovation. So, if you're a board, you're a senior leader, you want to do something. How are you trying to discover bias in your models? Some companies are setting up ethics councils. That could be another place to start.

Shrier

So, the Ethics Council discussion was hot for a while and then periodically they've been getting disbanded. More recently,

Saviano

Big tech

Shrier

Big tech. Correct. And that is that's a mistake that goes beyond the scope of today's discussion. But something that I will say is, you know, if you frame it to boards around risk and opportunity, So sadly, perhaps we live more in a shareholder than stakeholder capitalism world.
So, you've got to communicate with boards in the language that they're used to hearing and in in in shareholder capitalism terms, there is now unacceptable risk to the company financially and to EPS. If you fail to take this seriously and elevate the discussion to the board level. And in on the opportunity side, you could see anywhere from a 2 to 10 times expansion in your topline if you adopt AI intelligently.

Saviano

And how boards deal with that. Some boards actually separate AI risk from growth and innovation. They have a separate risk committee for us. We can debate whether that's good or not. Maybe it's good for one board and not for the next,

Shrier

But at least they have to have both of those discussions at the board level.

Saviano

They do. And whether it happens through a committee, you could have a subcommittee to your audit committee that has those responsibilities. So, there's many, many different ways that you could do it. But the importance is, is that that you have to have those conversations. It can't be ad hoc.

Shrier

But much like how boards were facing cybersecurity issues three years ago, absolutely a good analogy. In fact, a strong analogy, except there's one added piece which is the revenue upside.
So, I was talking to this company that does about 11 billion of turnover revenue in Europe, and he's looking at $750 million of incremental EBITDA from AI. A big chunk of that is being driven by revenue growth rather than by just cost savings. And so, for him, that's you know, a 25% increase in EBIDTA.

Saviano

And it's not just about the risks as you as you highlight its opportunities. And it's not just about new products, it's about entirely new business models. And that's been a lot of fun talking to companies about how could you use data as an asset or how could how could this be a catalyst to adopt an entirely new AI driven business model for the company?

Shrier

That's pretty cool. Whole new models of management emerge and become possible with AI.

Saviano

Okay, David, we have as a regular feature, as if this hasn't been fun enough, a regular feature of Better Innovation. We close the interviews with some quick questions and quick answers.

Shrier

Okay, so give me with the Lightning Round.

Saviano

Are you up for it? Let's do it. This is the lightning round portion. Okay, you've already answered this question. I'm going to ask it again. What's a book that has greatly impacted you? Well, you mentioned a book early that you really appreciate it, but I'll ask it this way. A book that has impacted you.

Shrier

Recently. I read a fascinating book called The Human Element, which talks about essentially the culture of what is needed in order to get innovation adopted, because ultimately, innovation is not a technology problem. It's a mindset shift problem.

Saviano

I kind of paused. I think that's the book. We had the author on the show. really? If my team is, that's great. But I don't. Before I answer that, because I think that David's shown those books, the human Element. Yeah, Yeah, it is. That was David's right, there were two authors, David And this was about resistance, right?
Yeah. You had over, over. okay. Yeah. I want to say. yeah, it was John David. Okay, Yeah, yeah, yeah. David That's amazing. We had the other David were going through a string of David's on the show. Sure. David's actually came on the show last year and talked about what we loved about the human element is that everybody in the world is focused on features and functions.
Yeah. And he a book about. That's okay, that's important. But focus on resistance and barriers and what are those like, what's the drag? What's the drag on a product? I came away, we had him on because I was blown away. I feel like everybody just has a natural gravitational pull to what's the next feature. But his approach is so practical, isn't it?

Shrier

Yeah, that's right. Well, it's also if you think about organizations and getting them to, let's say, adopt AI or adopt any kind of new technology, you know, 20% of the people are going to leap on it. 50% are going to move slower than you expected, and 30% of the people are going to make a lot of noise and fight it.
And that 30% becomes drag on the boat that pulls everyone else back. So, how do you enlist these people and bring them along with you? Excellent. Yeah, it was a great episode. We'd love to have a team on.

Saviano

Okay, Excellent. What piece of advice would you give to a younger version? Not that the current version is that old, but younger version of yourself.

Shrier

Trust in yourself to pursue your passions fully.

Saviano

So funny. We've had a bunch of guests coming in, essentially boiling down that question to I'd give my younger self a break, something to that effect. There's a little bit of that in your answer to.

Shrier

Yeah, a little bit. I don't know. I mean, you know, you could argue that any high performer is going to be hard on themselves, but I honestly think the, the converse of that is trusting in yourself more.
Right. I should have had more confidence in myself at various points.

Saviano

But that's good advice. It's really good advice, especially relevant now as we may be seeing people pivoting in their careers. Okay. Last question. What areas or industries are ripe for innovation in the next 3 to 5 years?

Shrier

I think we covered several of them in the podcast here, but certainly management, consulting, tax audit and advisory legal profession, investment banking, I think beyond that
I'm looking at a number of other industries where we're applying AI as platforms for change on $1,000,000,000,000 problem. And so that includes manufacturing and chemical, that includes pharmaceutical, that includes, that does include financial services and investment and asset management. You know, the thing about what we're seeing now is that so many industries are coming under pressure and have opportunity for change because of AI.
So, in the eighties we might have said manufacturing no brainer. We're going to automate all the car factories and the steel factories. Now it's like almost everything.

Saviano

I think we mentioned probably seven or eight different sectors just on the show today. Incredible. And that's what makes it so exciting. David This was another of our great conversation. We happened to have a couple of microphones to pick it up today but appreciate you coming in the studio and I really enjoy the conversations that we have about these topics. Thank you so much.

Shrier

Always a pleasure, Jeff. Thanks for having me.

Presenters

Jeffrey Saviano

EY Global Emerging Technology Strategy & Governance Leader

Podcast

Season 7, Episode 5

Duration

1h 12m 43s

In this series

EY refers to the global organization, and may refer to one or more, of the member firms of Ernst & Young Global Limited, each of which is a separate legal entity. Ernst & Young Global Limited, a UK company limited by guarantee, does not provide services to clients.