The Munk Debates Podcast - Friday Focus: AI Debate Recap – Titanic Mistake

Episode Date: June 23, 2023

Friday Focus provides listeners with a focused, half-hour masterclass on the big issues, events and trends driving the news and current events. The show features Janice Gross Stein, the founding director of the Munk School of Global Affairs and bestselling author, in conversation with Rudyard Griffiths, Chair and moderator of the Munk Debates. The following is a sample of the Munk Debates’ weekly current affairs podcast, Friday Focus.   On this week’s edition of the Friday Focus podcast, Janice and Rudyard start the show with a recap of last night’s mainstage Munk Debate on Artificial Intelligence. What did we learn from the debate? And, which arguments, “pro” and “con” on AI being an existential threat, had the biggest impact on the course of the debate and the audience? On the back half of the show, exclusively for Munk donors, the conversation turns to this week’s disaster at the Titanic that saw a catastrophic failure of an experimental, private submarine and the deaths of five people. What does this accident say about today’s culture of innovation, risk-taking and regulation? And how should governments respond, at great public cost, to the failures of for-profit risk-takers? This podcast is a project of the Munk Debates, a Canadian charitable organization dedicated to fostering civil and substantive public dialogue. More information at www.munkdebates.com.Become a Munk Donor ($50 annually) to get 72-hour advanced access to the full length editions of Friday Focus and Munk Dialogues. Go to www.munkdebates.com to sign up. Hosted on Acast. See acast.com/privacy for more information.

The following is a complimentary excerpt of this week's edition of the Friday Focus podcast by The Monk Debates. To access full-length editions of each and every episode, along with all kinds of great additional benefits and perks, become a donor to the Monk debates. You can do that for as little as $25 a year, and you'll receive each and every year 50 Friday Focus episodes at full length. It's all available right now on our website in just a few. you simple clicks, triple W the monk debates.com. Look for the Friday focus option in our navigation bar, the top right of the website. Make your donation and we will send you each and every Friday a link to listen to the full-length edition of this program. Thanks in advance for your generous contribution. Hello, Monk listeners. Rudyard Griffiths, the executive director of the Monk Debates.

Welcome to this, our regular Friday Focus podcast. This is the program where we dig into the big issues and ideas in the news each week with Janice Gross Stein, the founding director of the Monk School of Global Affairs, an internationally renowned scholar, and author. Janice, you and I were up late last night and now we're at it early Friday morning recording this podcast. Tell me, what did you think of the debate? We sure were. And I frankly, leave aside the content for a moment. We'll come to that. I was dazzled by these scientists, Richard. Let me just remind listeners, this was the Monk debate on AI in case you missed a thousand emails that we sent to.

My apologies in advance. We hosted four big names on AI in front of us, sold audience, 3,000 people at Roy Thompson Hall. Yan Lacoon, the chief, what were you called him, Jess, the chief kind of scientist for meta. Yeah, yeah. And a leader, really, in his own. own right. As you said last night, a researcher, a professor in his own right. That's why he went to Matt. Yeah. And Yasha Benjo, a friend of yours, who's leading probably in Canada, the most important international voice on AI. Rounding that out, Max Tegmark from MIT, a kind of charismatic,

sweet, tall guy, handsome, was good on his feet on the debate stage. And then Melanie Mitchell, who we can talk about really maybe the surprise of the debate, very forceful arguments to say AI is not an existential risk. She's from the Santa Fe Institute and has written a couple really big bestselling books on AI. You know, just to set the scene for all our listeners who weren't there last night before we go out of the content, there were these scientists didn't fit the stereotype at all. three men showed up in jeans. You know, when have we had on the monk men stage? Debate dominated by genes, frankly, I can't remember. I don't think we've had.

And secondly, these people are serious scientists, all of them, who have made groundbreaking discoveries along Jeffrey. And have created the field of deep neural networks and AI, remarkably articulate and frankly debated a subject which was not directly related to science. They weren't telling you how they put code in and what comes out once they put the code in. They were talking. The debate really was about the broader social consequences of AI and what that could mean for all of humanity. and they were articulate, focused, and to the bigger, the biggest surprise to me, even funny at times.

So for me, they broke the mold of the character of the scientists. And I think you're right. Melanie Mitchell stormed the stage is all I can say. She came out forceful. She stood up against two titans in her field. A bit of man. a bit of mansplaining too on the stage the guys were kind of trying to talk over her and she punched right back and i'd love to see that it was great she said just a minute let me finish and boy they all

quieted down and she gave no ground so i think just a terrific debate and i would just encourage all our listeners if you weren't there last night run your wedding will it go up so people can watch yeah well it's up right now on the website so if you are a monk donor if you're either a monk donor if you're a supporter, a curator, or you've given us some money to listen to this podcast, which you really appreciate. You can access it free right now on our website, the entire debate in glorious high definition. Just go on to the website, use the login function and your password and email. And boom, you got it. If you're not a donor, you can get not only this debate and the next live stream in the fall, we give you two a year free and archived. You also get the back half of Friday

focus where we're going to have a conversation today about a sunk submersible that's been dominating the news so join us for the back half of the show grab a monk membership a donor contribution as little as 25 dollars it's all yours right now in our website triple w monk debates.com jenis let's i'm curious what of all the different arguments in debate pro and con AI being an existential risk, which resonated most with you? If there was one argument that you took away from that evening and thought, okay, that's the argument that pushes me into the pro camp, that there is an existential risk here or the con camp. No, there isn't. This is a technology that we can manage in the future. Which was it? I think it was Melody Mitchell who said,

there's no demonstration of how this is an existential risk. There's no, in a sense, you can't demonstrate, you can have evidence that something in the future is an existential risk. But people who claim it is an existential risk, Rudyard, want the research slowed or stopped. And we already know, and she made this point, Oh, in a relentless way over and over, there are present harms and present goods. And she argued so forcefully that focusing on a far out existential risk that a superhuman

intelligence is going to emerge from all of this and control all of us. Because that's really what the argument is. It's about control is a misdirection of our time and effort. We need to focus on the here and now what the technology is. doing and invest in sensible regulation as we have as she said in all other technologies before so jess what do you think of this idea though which came up in the debate of a precautionary principle that there are other examples of human caused events let's say climate change is the is the key one where many people rang the alarm bells decades ago and said look you're emitting all this carbon into the atmosphere

it's going to have big consequences and the big industries that controlled high hydrocarbons said, no, no, no, we have our own think tanks that we've bought off. Here's all this research that says this isn't a threat. Don't worry, it's not existential. Nothing to see her folks move on. Where are we at two, three decades later? We're now suddenly in a climate crisis. Our forests are on fire, record heat waves around the world, record ocean warming.

Isn't that an example of how, if we had understood the existential risk of climate earlier, we would have acted and prepared an absent existential risk. We're just going to let these corporations do what they always do, which is race ahead, race ahead and frankly create justifications why we should be racing ahead and not being more prudential. So let's talk a little psychology for a minute here, Redd. There is a long human history of people having short time horizons and focusing on your term. As a species, we are lousy savers, as you know. To give you a concrete example, people don't save for their retirement because it's too far off. The time you should be saving for your retirement is when you're young. It's a real push to get people to do that because

we discount. Here's the weakness, I think, and why I don't believe that we're going to do much until there's more to worry about and it's more visible. The pro side who said, yeah, this is an existential risk, and Joshua Benjillo and Jeffrey Hinton and Max are at the forefront of this. Well, what's the risk? Tell us, give us concrete, vivid examples of the risk. I don't think either of the two debaters did very well at that. They didn't make it vivid live.

They didn't say some superhumanoid robot. It's going to be telling you what the next debate subject is, right? You have to make this come alive for the public. So sure that we have precautionary principles and we can do that for AI now, we could create a national registry again. And just like we have for drugs, right, for new drugs. And you go and you tell the registrar, what the harms are and what the goods are and how are you going to mitigate the harms.

But right here, you'd be the first to say that some of that regulation drives you up the wall and slows down innovation and makes the process cumbersome. So you have to be really convinced that slowing everything down because it throws sand in the gears. But in this case, I would support it, Janice, because, you know, as Max Tegmark and you mentioned the FDA or Health Canada. In some ways, these, these are like a regime that could be thought of in terms of the, whatever you call it, the biotech industry. Like you can't just go and create a pill for disease or condition X and start selling it in supermarkets. You really should take this technology, which is going to be so ubiquitous, will be very powerful. Well, as you say,

have upsides, but also downsides. And I think, it's exactly the right place to have some kind of national regulator or somebody who says, you know what, we're going to look under the hood here. Maybe we're going to do this with you privately. We're not going to expose your intellectual property to your competitors. But we're going to have a full exam of what this algorithm, what these learning models do, what goals you've set for them. And we're going to have experts assess this. I think this is a no-brainer. I think that's where we are right now. I think that's where we're going. because there's a real, there's a global push now for regulation.

And I'm not opposed to regulation, Roger. But what, but how nimble, how flexible? It's going to take two years, three years, this close examination of under the hood. Because frankly, and here's a really interesting issue that it didn't come out as sharply as it should have last night. The people at the forefront of developing these new technologies don't understand what's under the hood themselves. So if you said to Joshua, tell me what's under the hood in your latest product, he can't do it. That's what's freaking these people out. That's what's freaked out, Jeff Hinton.

That's what's freaked out Joshua. That's why they've done this 180-degree turn. So saying a regulator look under the hood, he can't, right? So what we call this technically an emergent property. Some will emerge in ways that we don't understand. Makes that kind of product-specific regulation really tough. Nevertheless, I think we're going there, but here's my cautionary note. Think back to COVID.

Think how infuriated you were that it took us six months longer to approve the first vaccine than it did in other countries because our regulators slowed it. Think how infuriated we were that simple off the show. shelf rapid antigen tests had to go through the health Canada regulator. And we couldn't get it on the shelf. A simple product like that. And we were behind the United Kingdom and the United States because we have our precautionary principles are so wired in favor of eliminating any risk. Agreed. But I think this is different, this is worse here than a vaccine, this technology. And as Yan LeCoon mentioned, you know, we didn't allow, because there was discussions about in the beginnings of the atomic age, about nuclear powered cars or rocket ships. We didn't go down that route because it's not too safe to have like 30,000 cars on a highway powered by nuclear reactors.

We just didn't do it. And I think that's how we have to think of this technology. We have to elevate it in our minds, maybe for some. it is an existential risk for others. It isn't, but we have to classify it, I think, differently until we, if we can, understand it. Let me try one other thing out on you before we go to break. And again, just encourage our listeners to, if a donor, grab your free version of the debate right now in our website. If you're not a donor, please consider supporting the monk debates.

And you get this entire 90-minute-plus exchange and beautiful HD video on-demand. any time and a live stream of our fall debate and the back half of Friday Focus. What a deal for $25. This is the argument that I thought was a bit new in the debate. And then I came away with thinking about more on the other side, the side in favor of existential risk, was not the doomsday scenario of the Terminator robots or the mad scientist in the lab who uses the AI to cause all kinds of havoc. It was this idea of Max Tegmark brought up of a slow,

erosion of agency through the power of this technology whereby individuals and corporations and governments that use AI will just perform better than those that don't. And there'll be a big first mover advantage and there'll be a lot of pressure to jump on the AI bandwagon to hand over human decision making to machines because machines will be less biased. They won't have all kinds of hangups or cognitive blinders. They've got trillions of data sets to input to analyze things. And the result of that could not be some like existential moment where the light bulb goes off. Instead, it could be this slow dimming of human agency where we wake up 2050, 100 years from now. And we're not making any decisions about anything because all of our agency has passed over to

these machines that are telling us what to eat, think, watch, how to create a carbon tax credit, how to design a more efficient microwave. I don't know. But that to me is a scary scenario that somehow to me gels with how I think humanity could interact and relate to this technology. And our quest for power, we give up power. There's a irony here that I think is deep and one that's worth mining. So you're going to be surprised to hear whether you have zero sympathy for that argument, frankly. It is, and it really stems from this profound human worry that we're going to lose control,

right, and that our own role. And frankly, freedom to choose, which is very important to all of us. But I think that's a deep misunderstanding of the history of technology over time. I could use that argument to say, oh my goodness. And Yonle Kuna thought was great on this last night, right? Every technological event is big, not small little steps forward, has done what you just described. It has empowered some and disempowered other. And the best example he gave last night was the printing press, right?

The church absolutely frantic about it. But why? because they control and all of a sudden the technology diffused other people could write you could have anti-clerical publications and he's not wrong it broke the monopoly power of the church so it never deprives people of agency let's spin but hold on jess this is intelligence it's not a physical processes like the printing press or an automobile this is a very specific It's not intelligent. What it's intel.

You and I have discussed in a previous podcast. I'm not saying it's human intelligence. I'm saying you and I, I think, agree that intelligence equals power. Yeah. And that this is power expressed not through the Newtonian, Brownian world of physical objects and things. It's power expressed through zeros and ones and ideas. Now, they may not be human ideas. they may not be ideas that we can comprehend, but the results of these ideas will be very powerful.

They will change the world. They will change us. That to me is new. That to me is potentially transformative. And that to me is potentially dangerous. And if we just conflate everything into saying, well, this is just like the past. There's nothing new under the sun. I don't know. I think Joshua Benjio and Max Tegmark, if they were here, would say that's, maybe not as dynamic and creative as the thinking that we need at this moment to address the situation. Let's just not use the word intelligence for one minute. And let's talk about advanced computational capability. So I ask my AI, which will be personal to me, 20 years from now for sure, should. Is this the right time for me to buy a house? And I feed in all my demographic data. and everything else. And it gives me an answer in one minute and I don't have to do any work.

But the agency doesn't go away from me. I still ask the question. It's my goal. But there is this sophisticated computational capability. That's what AIs are. You know, it's really interesting after the debate. Of course, there was a big debate in the hall outside. Somebody said it, let's ban the word, intelligence in this. Because it's so misleading. If you take that word out and you say, super computational capability. You wouldn't worry about agency. But that's all this is. Super advanced computational capability.

Just one final whack at this. I think the cautionary tale, Janice's social media. So the analog to artificial intelligence, the kind of beta version, the simplistic version, which we created these very powerful computational models, machine learning, to get people to stay on websites,

watch ads, and click links. And these algorithms have become incredibly good at doing this, at holding and commanding our attention. And I don't think anyone can make an argument that we have lost some agency to these algorithms. We are fooled into thinking that we're engaging in, you know, free, conscious, rational choice. But how many of us, Janice, know somebody who is completely. gone down the rabbit hole, who has lost their bearings on terms of what is real and what is not, not because the machine wanted them to go insane, but the machine made them insane by being so good at pursuing a goal that it was tasked to pursue without an awareness of the casualties,

the externalities of going after that goal. I think there's a risk here that AI could be social media on steroids. And agency has been a key issue, a key concern in the social media debate, and I fear it mapping over to AI. Yeah. So I understand that. And I have a very little use for social media the way it's evolved. I think it's toxic. And it's causing tremendous social harm. There's no question about that. That's a regulator problem. Look what's happening in Canada. Maybe we can end here. Facebook is going to cut its newsfeed off for Canadians because, frankly, a very small regulatory step that the Canadian government is taking to support the newspaper industry, right? So this is an older story, frankly. If governments stepped up and said you take out these addictive algorithms which go up the scale of toxicity, frankly, you take them out or you don't get access to our digital space,

the problem you just put on the table would be solved. But we're not solving it because these corporations are some of the most powerful entities in the world. Well, there's a lot of asymmetry here. And I don't know. We haven't solved it yet for social media. We know all the harms.

We're a decade or more into that experiment. Like tobacco, like tobacco, right? How many decades were we into it? Once the surgeon general said, This kills you, literally kills you. How many years did it take? So here's the bigger problem you're raising, right? Smart regulation is hard to do and it always lags.

But we get there. I'm a relentless optimist. Well, let's go on to our next topic. We'll take a short break. We're going to discuss what's been dominating news all week. This sub-accident, Ocean Gate, failed fatal dive on the Titanic. I've got a hot take on this. I want Janice to react to it.

We'll get into that with our monk donors right after this break. Thanks for listening to this excerpt of the Friday Focus podcast. To get full-length editions of each and every episode of this program, simply go to our website, triple-w, the monk debates.com. Click on the Friday Focus tab in our navigation on the top right of the site. Make a donation as little as $25 a month. year of 50 cents an episode and we'll send you not only the full-length editions of each and every

Friday Focus podcast, but all kinds of special offers, perks, access to events, and additional content. Again, you can do that right now by becoming a donor to the Monk Debates at Triple W. Monk Debates, MUNK, DebateswithanS.com.