The Pension Blueprint podcast video transcript 

Episode 4: AI and the future of your pension: Smarter tech, stronger futures

Celine Chiovitti (VO): Hello, everyone, and welcome back to The Pension Blueprint. I'm Celine Chiovitti, Chief Pension Officer at OMERS, and I have the pleasure of being your host this season.

This week, we're talking about two things that, on the surface, might seem entirely unrelated: artificial intelligence and your pension. The topic of AI is everywhere, and its advancements as a technology over the past few years have been staggering. It's brought new opportunities to invest and to optimize working systems. At OMERS, we're cautiously taking advantage of both.

On this episode, I'm joined by two in-house experts: Ryan Contant, Senior Vice President, Pension Services and Technology at OMERS, and Henry Gladwyn, Managing Director, Ventures at OMERS. We'll talk about how innovation in the AI space is impacting your pension, as well as what you can do in your own lives to lean into AI, and how to protect yourself from possible scams. So let's get into it.

Celine: I'm really excited to have two innovative, savvy, and business-minded leaders here with me today to talk about all things AI—the good, the bad, and the ugly. Welcome, Henry Gladwyn and Ryan Contant, to The Pension Blueprint.

Why don't I start with you, Henry? Tell us a little bit about yourself and your role with OMERS.

Henry Gladwyn: Well, thank you very ever so much for having me, Celine. I'm in the Ventures team based out of New York. I originally joined OMERS six years ago in the London office, also doing Ventures. Ventures for us means early stage technology investments, principally in business, business software applications. My focus is particularly on what we call vertical AI, so AI applications made for specific industries or specific businesses.

Celine: Great, welcome to The Pension Blueprint, Henry.

Henry: Thank you for having me.

Ryan Contant: I'm so excited to be here, Celine, and thanks for having me. My name is Ryan Contant and I've been with OMERS for coming up on 15 years now. I am responsible for our Products and Technology team, which sits in Pensions. We're about 170 individuals strong, and our whole mandate is to provide the best pension experience to our members through technology, of course AI being a major part of that.

Celine: So Henry, you work for the Ventures team. Can you talk a little bit about how Ventures fits into the overall OMERS investment strategy? And then when you think about Ventures in itself, what role does AI play in it?

Henry: Absolutely. So Ventures is slightly under 1% of OMERS' total net assets. So we're a very small part of the overall OMERS book and that's reflective of—we're probably the highest risk end of the spectrum, and I hope obviously highest return as well. Within Ventures as well, AI is a fairly new technology. It's been developed for a long time, but as an investible technology, it's three or four years old.

So within that 1%, it's probably at this point, I would say, somewhere between 10 and 20% of our investment activity has gone into AI—what we think of as AI native businesses. We expect that to grow over time, but we also don't expect normal software, normal computation to go away. So I don't think there's ever any point where Ventures is 100% AI, but I do think there probably is a point where Ventures, and frankly, most of the other groups, are significantly informed by AI and using it—every business using it on a day-to-day basis.

Celine: So Henry, I'm going to go to you and just ask you, what does AI mean to you from a technology space? How does it add anything to OMERS?

Henry: Yeah, a very good question and one that I ponder periodically. I think the way that we think about AI as distinct from other types of software or other types of technology is this concept of non-deterministic compute. Most software is entirely what we think of as deterministic. You design an algorithm, you run the algorithm, and it does exactly the same thing every single time, and you hope that it's right. It's been great, it's been very powerful, it's been very useful.

But what AI, or non-deterministic compute, brings is, I think, three things. Firstly, because it's non-deterministic—so it's able to make guesses, try different ideas, rather than saying "if this, then that," it's saying "if this, then most likely outcome"—you can address a lot more problems. The world is not, for the most part, deterministic, although we've done a very good job at twisting computers to being able to do it, so you can do more things with it.

Secondly, it's more versatile. If you've, as I'm sure you have, Celine, used something like Excel—a very powerful program in the hands of someone who really knows how to use it. But if you get even a little bit wrong, the whole thing will go wrong. I'm sure we've all seen broken formulas and things like that. Whereas with non-deterministic or iterative compute, you can really talk with the program. So it's a much more versatile experience. You can get more from it if you know how to use it, but even if you don't know how to use it, it becomes much easier to use.

Third, and perhaps most excitingly, it gives the opportunity, perhaps for the first time, of a computer that can find things for us that we didn't know ourselves before. We've seen that in things like protein folding. We saw it with the Go games about ten years ago where, for the first time ever, we have a computer program that has effectively discovered new, or at least novel, knowledge for us. Whilst those have been small examples so far and they've not been great breakthroughs that changed our lives, that, I think, is for us probably the most exciting possibility—that we now have a tool that can find things for us.

Celine: So I'm going to just push you on that a little bit, because there's so much hype around AI. When you're actually looking to source an investment in this space, what would you be looking for in sort of an AI startup?

Henry: Yeah, so you're absolutely right, there's a huge amount of hype around the space, and technology investing is by its nature cyclical. We have seen in the past these big waves, and not every new technology results in a boom-bust wave, but most of the most important ones do. So you have to be careful about how we select things and careful about maintaining a truly long-term view of these things and not sort of going too hard, too early as it were.

So we are not necessarily looking for the crazy, insane growth, because I think that often points to growth that may not be long-term sticky. What we are looking to see is that people have adopted this technology, they've changed the way that they're doing their business, and that it's become seriously integrated into the way that they do their business—such that if you were to take it away from them, it would be difficult.

So I think the classic print that we are trying to make here is to email. Once people had email in their business, it was very— I mean, we all hate it, but it's very hard to go back from it. I don't think that it's necessary we'll be able to get hold of something as powerful as that, but that's the pattern we're looking to match to.

Celine: I love, because what I'm hearing from you, and I think there is this fear that AI is going to take over the world and take away all of these human jobs. What I'm hearing from you, I think, is that you're actually looking for ways to help optimize the work that we're already doing, sort of in your world, in the investment space. So how do you sort of take things, rely on AI to help people integrate that into their daily jobs and get even stronger at what they're doing?

Henry: I think that's exactly right. The people who build these types of technologies are, by their nature, enthusiasts. That's why they do it. In the history of AI or the history of computers generally, there have been a lot of claims that have turned out to be overblown. If you go back to the 1950s when we were first talking about this, people thought it was going to be done. You know, it'll be done within the— And if you go and speak to some of the leading AI experts today, they think it's going to be done really soon. I suspect, 50 years from now, people will think it's going to be done really soon. But that doesn't take away from the fact that it is a very useful technology in exactly the way that you say, which is helping people do their jobs better. That's really what we are looking for—companies that can help you do your job better.

Celine: So on that note, I'm going to turn to my friend, Ryan. You run a big business, and the work that you and your team does really is very human-centric, providing services to OMERS members—over 640,000 plan members. Your team is responsible for the pension platform, so all the technology components and modernizing it. As a business person, taking the hype away and just listening to everything that Henry has said, have you found any actual use cases where you've been able to deploy AI in a way that's been able to assist your team?

Ryan: We are always looking for use cases, because to us, and to me specifically, AI is just a tool in the toolbox. It is not the tool in the toolbox, it is a tool. So for us, we want to find use cases that fit that model. I think you have to think about a couple of things. The first is it's not perfect. We know that there's an error rate to AI, somewhere around the 90 to 95% in terms of the use case. For us, it has to be a use case that is okay with that.

So if you think about when a member calls into the call center, there are a lot of different things that are happening behind the scenes, including somebody taking notes. That is an excellent use case, because we know AI is great at both summary and analysis overall. We can take that data and information, and instead of a user actually writing the notes, which takes several minutes over time, they can actually focus on the conversation to provide the best experience to our members and let the tools take care of the summarization, and really provide just the next details in line for when they call in next.

Celine: That's great.

Ryan: So again, finding a way to help optimize, not remove. Email is a really great example of a new technology that emerged and we were all resistant for a little while, and now, of course, everybody can't live without email.

Celine: When you think of that, Ryan, in your space in running an operation, how do you embrace new technology? How do you get your teams to start to adapt? And does that happen overnight?

Ryan: So the answer is no, it doesn't happen overnight, Celine. I think what we see, at least on the floor, is our team loves technology. They come to me with many, many different ideas about what can we do, what can we try, AI being one of them, of course. For us, we are an agile team. We will always try new things, and if it doesn't work, we're happy to pull the plug. But the truth is, being technology first and providing the best experience to our members, we have to play with some of the new tooling. With that, there will be successes and there will be failures. But the real point that I want to hammer home is that we will adjust if it's not working for us.

Celine: That's great. Thank you. Thank you.

Do you—I'm going to go back to you, Henry, for a moment. What are your thoughts on whether this is going to go away one day, calm down one day, or should we all be really, really paying attention to this space?

Henry: Well, obviously, talking my book here somewhat, Celine, but I think the answer is both. It's not a totally satisfying answer, I know, but the pattern you tend to see in technology is this: over promise, under deliver, but eventually deliver more than we could ever have hoped. That's the pattern. The classic example of that is the dot-com boom, which we would have all experienced and many of the members will recall—a technology that was clearly a powerful and useful technology. For various reasons, it was overhyped. There was then a crash where it was underhyped, because if we now look 10, well, 20 years later—more than 20 years later—8 out of the 10 largest companies in the world are digital internet companies. As long-term investors, we have to think in that way: there is no idea that is so good that it can't be overhyped, but I think this really is a very powerful technology, and we need to think about it as a very powerful technology that, in 10 years' time, 20 years' time, will have brought real change in some of the ways that we've already talked about. But that does not mean that it will be all up from here, and we should expect significant volatility, at least in part because we don't fully understand the technology, we don't fully understand these applications, and we should investigate into that in a prudent and sensible manner, remembering that there will probably be a boom, and where there are booms, inevitably, a bust.

Celine: Don't be a bust. I think those are really great words for all of us to live by: pay attention to it, continue to move forward, but don't, again, ignore the hype in the space.

Henry: And I think the flip side of that is, in as much as you ignore the hype, you then also ignore the sort of anti-hype. I think one of the mistakes a lot of people made post dot-com was that they sort of wiped their brow and said, well, that's all over, you know? We don't need to pay attention to that. That was the wrong conclusion to draw, and I fear there may be a risk of doing the same thing.

Celine: So Ryan, when you hear that and you're thinking about the future—the future of OMERS, the future of pensions, and the future of members—our youngest member is 14. I know one of the things we've been talking about is how do we innovate for that next generation? We run a contact center. You talked about that. I look at my daughter in her mid-twenties, and I often say her idea of picking up the phone and talking to an actual live human between nine to five, I think, makes her shake because she's just not in her comfort zone. How are we preparing, what are some of the things we're thinking about for five years down the road?

Ryan: Yeah, I think we're approaching it with optimistic caution, is how I would say it. The reality is when you use a chatbot today through any platform, you're more than likely going to be disappointed, and I think that is not an experience that we're ready to enable for our members. What I mean is really just allowing them to communicate with the AI directly. So we prefer, again, to have a human in the middle, using it as an assist tool versus full-on going. Now, if you think about 10 years from now, the reality is our main focus is optimization and efficiency. I think what the members and our membership will see over the course of the time is that we will be able to pay faster, sooner, more accurately as we go through the modernization of our systems, leveraging AI where it makes sense.

Celine: What are some of the actual tangible things that I can do today to help myself from a change perspective? What should I be paying attention to? What should I be educated on? How do I ignore the hype but really focus on some real tangible skills that will help me in my role today and into the future?

Maybe that's open to both of you, but I'll start with you.

Ryan: Do you want me to take a crack at it? So I would say the first thing that everybody needs to do, or should do, is play with it. I think it is very easy, it's accessible. You don't have to pay, although there are paid options, but go and understand what all of these LLMs could actually do and what they're good at and what they're not good at. And I think that's the first thing.

What I'm seeing on a daily basis is people using it for writing. So, we talked about emails earlier, we talked about—
Copilot.

Exactly.

So Copilot can help put a professional swagger on your emails, and it's just, you know, spell check times 10 or 15. We're seeing that all of the time, which is really, really helping and making for some really great conversations around the globe.

But I think the reality is people need to understand that it is not perfect. It is at its infancy, is what I would say. And again, I will say I'm an optimistic skeptic, and I think that the reality is you just have to get in and play with it and see that it's not going to change the world as expected today.

Celine: And I think I love all of that. And I think when you go back to using it as a tool, because I think to your point, Copilot and all of those, you can get access to almost any piece of information. It helps you perfect a sentence, a paragraph, a story, which I think is phenomenal. What I worry a little bit about, what I would caution around is you still need to think. Now I've written a paper, but now I have to be able to answer questions and show the depth of knowledge in that space. And so that's sort of my own—maybe that's a bit of my old school—

No, I think it's—

Coming out, but that is my, yeah. What do you think about that, Henry?

Henry: I completely agree with both points. I think what I have found with it is it's a great tool for making you better at things you're already good at. Writing is a great example. So everyone who has to answer emails is pretty good at answering emails. This would make you more polished, quicker. Everyone who has to do research as part of their work is probably pretty good at doing research; this will help you do it.

And then on the other end of the spectrum, I've found it can help me get up to speed with things that I don't know very much about. For example, I'm not a particularly good coder, I wanted to do some coding, and it pretty much did it for me. But that point of it can do it for you—it may not be very good. Not just that you won't be able to answer questions on it, but sometimes you can ask it to write an essay or something—people do this for school all the time—and then read it. You know, this is actually not very high quality. So using it genuinely as a tool, rather than as something that does your work for you, is a really crucial thing.

The one thing that I've been surprised by is when you are trying to learn new things with it, which it can absolutely help with, you do have to keep checking that it's right. You do have to keep checking that—most of these programs are optimized towards just making you feel good. So you do occasionally have to pause and say, why am I wrong about this? Or what else should I be considering? But otherwise, I completely agree. You should view it as a tool. You view it as a tool that you can play with. And then, crucially, because most people aren't playing with it, you can put yourself into the top 10% of AI people in the world just by giving it a go.

By getting comfortable with it and going, playing with it.

Celine: So on the other side of that, how do we protect ourselves? I keep hearing about all of these scams—how do you fact check? And, again, going back and thinking of OMERS plan members, 200,000 of them are retirees. I do think that, you know, the negative side is we are more vulnerable to scams and cyber breaches and things like that. Any points or any tips for our members on what they can do to protect themselves in this new technology space? Do you want to go?

Ryan: So I think just like everything in terms of technology, people are targeted, whether it's through AI or any other phishing scam, and it's just about being smart and thinking about where is something coming from. The reality is the emails are very well written, and it doesn't sound like somebody that is struggling with the English language, it sounds great. So I think the reality is whether you're checking your email—if it's a call, double check. Always go to a website. Don't just log in and click links, go to the website and log into your account directly. But it's just about the education piece, knowing that there are bad actors out there, and we have to be cautious.

And they prey on your vulnerabilities, because I've sort of heard the horror stories about, you know, I've gotten a call from my, you know, I thought it was my grandson who was in a different country and they asked for money, and I wouldn't normally give it, but I did it because of this. And so I think it's that they—it does prey on where you're most vulnerable and feels very real.

And I think we are all, I think, used now to spam and phishing, and most people are pretty—it's those emerging ones of, it's voice now, and it'll be video, I think, fairly soon, where you just, I think you just need to be aware and know your grandson well enough.

He doesn't speak that way.

Coordinate a code word, yeah.

Henry: I mean it is—you joke, but quite possibly things like, I think when you think about deepfakes and so on, either technologically engineered code words or within family groups or within that kind of thing, I think is probably where we end up going to avoid that kind of stuff.

Celine: I think those are the conversations we need to have. So, summarizing the past two points: let's not be scared of AI, let's leverage it, let's lean in, let's play with it—especially those in my generation that aren't necessarily as comfortable as, you know, people growing up today; this is just common, this is what they're used to. And at the same time also, be cautious and do the fact checking and proceed with caution—there are some real actors out there, it does feel very real, and it is easy to fall and to prey on that.

So I think really good words of advice. So, I mean, you must see everything.

I would love to just be—I'd love to live a day in your life for one day. But if we're looking to 2030, 2035, walk us through what this could look like. What does the world look like in that space? What are some real innovations you've seen?

Henry: Leave yourself immediately hostage to fortune with any statement here. I think that the things that we are seeing that are particularly exciting at the moment in terms of applications of this technology—I mentioned just now video—that's only just beginning, with some big releases this year in terms of almost studio quality. It's very limited use cases at the moment, but you can see where it goes. There are obviously dangerous uses of that, but to the point we are making about playing, at the moment, you can go and play and you can generate emails, you can generate documents. I think fairly soon, creative young people, particularly those who are sort of born to it as it were, will be going and playing and making video games and making videos, and I think that's exciting.

In terms of things that are adding, though, to the economic growth story, the one that is coming down the pipe, I think, is what is firstly, agents. So that is AI-generated programs that do more than simply give an output—they can interact in the world in some way and give more outputs. So you might, for example, we've mentioned here talking to chat agents in a contact center. Not very good at the moment, but again, you can see the direction these things are going. And then what are being called now embodied agents—so taking that ability to interact with the world, to a certain extent reason, and putting it into robotics.

I would think of the things that we are seeing that seem to have the most potential outside of the realm of computers or traditional technology, it's the future of either humanoid robotics or some form of advanced manufacturing powered by robotics that I think is particularly exciting.

Celine: So you think that's like five years away, 10 years away?

Henry: So I mean, I think we're going to see examples of it. It's a funny thing about technology, isn't it? There's stuff that 10 years ago, if we told ourselves that it would be happening—like the Waymo cars—they are self-driving their way around the streets of San Francisco and other US cities right now. I think 10 years ago, if you'd said, oh, that's going to happen, people would be like, really? So I think more of that. I would expect that within 10 years, we are seeing humanoid robotics in some form. And I've now made myself completely hostage to fortune.

Celine: Maybe we'll have a season 13 of The Pension Blueprint and invite you back.

Henry: Hopefully.

Celine: Okay, I've been ending our conversations with a series of fast five questions, and I'm going to rotate between the two. We have fun questions. Henry, first thing that you do in the morning?

Henry: First thing I do in the morning—oh, we're meant to be fast, aren't we? No, I'm just trying to—this morning, I sort of turned my alarm off and went back to sleep.

Celine: Well, there you go.

Henry: In theory, I get up and go for a run. This morning, I did not do that.

Celine: Either snooze or go for a run.

Henry: More often snooze than run, sadly, but yes.

Celine: Ryan, last thing you do before you go to bed?

Ryan: Kiss my wife.

Celine: Ah, very nice.

Henry: Prepared, Ryan.

Yeah, that was good.

That was fast.

Celine: Henry, best restaurant? What's your go-to restaurant in New York and what's your order at that restaurant?

So I'm an absolute pizza fiend, and so there is a pizza restaurant called Roberta's, which is very close to my apartment in Williamsburg. My go-to order is a pepperoni pizza. Hence the running.

Celine: Sounds good. Same question to you, Ryan, in Toronto.

Ryan: For me, it's any type of barbecue. I'm a lover of brisket and all smoked meats. I would say restaurants—they're a toss-up for me. Whatever's on the menu, I'm happy to go.

Celine: But you're ordering barbecue?

Ryan: Always.

Celine: Okay.

Ryan: Always.

Celine: Good to know.

Celine: And then the same question for both of you. You have 48 hours to go anywhere you want, excluding travel time. Where are you going, what are you doing, and who are you taking with you? Maybe I will start with Ryan this time.

Ryan: So for us, we're Disney lovers, so I would go to Disney, probably in Tokyo, with the family.

Celine: Fun.

Ryan: Disney in Tokyo.

I would go to Zermatt in Switzerland, also with my family. I think it's humorous there. I think it is sort of Alpine Disneyland.

It sounds amazing.

Celine: See, part of this is I'm getting travel ideas for—

Ryan: We'll see you at Disney.

Celine: There you go. Thank you both for being on The Pension Blueprint today.