[MUSIC PLAYING]
Should we be scared about artificial intelligence? Are the robots coming for us? We touch on all of this and more on Data Nation from MIT's Institute of Data Systems and Society. I'm Liberty Vittert. And today, my co-host, Munther Dahlet, the founding director of MIT's Institute for Data Systems and Society, and I are speaking with Devavrat Shah. As an MIT professor and tech entrepreneur, Devavrat Shah has seen firsthand how AI tools can impact research, business, and careers.
While some have dire warnings about the scale of harm AI can cause, Shah is optimistic. He joins the Data Nation Podcast to dispel some doom and gloom, unpack ways that people are already using AI to make changes for the better and to examine how future benefits can emerge from regulation and education. Let's start this discussion of the future of AI with, I guess, a personal question. And that would be overall, what is one word that would describe your feeling about the future of AI and why?
Are you hesitant? Are you fearful? Are you excited? Are you apprehensive? What would be one word that describes your feeling? Very optimistic. That was not what I was expecting. I love that. What gives you that optimistic feeling? I feel that there's one part where everybody is afraid of losing jobs. There would be doomsday. My computers will be hacked, attacks would happen, and all that. It all sounds fantastic.
However, what I feel is that really the biggest and largest value would come is just the way washing machines changed, electricity changed, having automated vehicles changed a lot of things for us, AI will play similar role. And AI is not everything, OK. AI is a small piece of a much bigger technological world out there that's there. And there is so much physical stuff that happens. AI is not going to change it.
Even if there is the greatest form of AI, there will be still amazing food that you and I will enjoy eating. So coming back, I feel very, very optimistic about AI is because I think in forms, it's been around for a while. I mean, we call it machine learning. We called it @IDSS data science. And now we call it AI. So just the way we have done a tremendous progress as both intellectual community as well as society at large, I think there's a lot of good things that will happen.
You read in the news, or you hear commentators, or you hear people creating this very fearful environment around AI and almost like we're in this AI revolution that's different than anything else we've ever seen before. You don't really feel like it's that way. This is the same as the cell phone or electricity being created or the Industrial Revolution. It's not something that is beyond our control and crazy and out of this world.
It's something that's controllable and gives better life than a worse life. Is that fair? So I think that there are two parts here. One is what you're saying is a tremendous change, and second is controllability part. So maybe let's dissect it in that way. I mean, cell phones, I have seen that change in my life.
There was 1999 when I came to this country and then 2002 when I had my first cell phone, which still remains to be my same cell phone number, has completely changed the way I do my day-to-day things. Electricity, I lived in a country where there were be days when-- actually, many days regularly, like weekly, there would be scheduled where parts of the day, there won't be electricity. So we knew how to live with electricity and without electricity. And that has changed everything.
There is a world which was before internet, before emails, and after internet and after emails. That has again changed things drastically. So in similar sense, I think AI would have value. Now, the thing about thinking of AI as sudden change, I think that's a mistaken. Because a lot of these things have been around for a while, is just that the popular attention, popular societal attention has come now a lot more especially in the recent times.
And maybe that's the reason why also we are having this conversation. But again, I think we've been with this for at least 20 years. I mean, since we started collecting large amounts of data, getting information out of it, and using it more and more, that's been what's been driving force. The very traditional view of AI has been whatever is artificial intelligence, there's a natural intelligence that is you and me.
Whatever we can do, machines should be able to do, and hopefully, at least in some aspects, they can do better. We can fold towels. Machine can fold towels, maybe a little faster. We can drive a car. Maybe machine can drive a car without falling asleep and things like that. So there's been automation, and there is a learning from data. A lot of modern AI has been about learning from data rather than thinking about automation as well as coupled. But I think that also needs to be thought about.
So in short, it is a great change. It's not been sudden that we should suddenly start worrying about controllability. If it has been controllability has been an issue, which it has been an issue, by the way, and we can come to that in a bit, but it's been an issue for at least 15 years. I mean, misinformation, that's been also very much about question of controllability of AI. Let me actually pick up and take the first part of your answer first.
And you are in-- your optimism, if I would say, comes from your unique position because you're an academician who has contributed to machine learning and AI, but also you're an entrepreneur, and you have used that in the business world. And I would say you've had a first hand experience seeing how AI and technology have had transformative effect on the business world.
And maybe you can comment a little bit about where do you see the biggest impacts in the business world and the job creation and all this optimism that you come with, which I also agree with, where do you see it mostly happening? And as opposed to the fear of losing jobs, is the job creation and the opportunities that are created? That's fantastic. I mean, I think maybe that's one of the key things we should start thinking about. So first is people worrying about jobs.
So it is no doubt that there will be jobs that will disappear. For example, before washing machine, people were washing clothes with hands, and those things disappeared. But then that created few jobs. One is people who build the washing machines, who managed it, maintained it, et cetera, et cetera. But that's not just the one thing, right. It's not just washing machine. It's a lot more than that.
People who have built AI solutions, who would be building AI solutions globally, there will be lots of positions there. There has always been this whole question of we do not have enough data scientists, right. Well, maybe with AI and all the tooling and all the development around, that question would hopefully disappear. But the way that would disappear is by us academics actually educating people how to use AI and then its ability to solve harder problems that have not been yet solved.
The thing about AI taking away experts will never happen. In the short term and a long term, there will be experts individuals who both have information that machines have not captured, who both have ability to do things that machines cannot do, and the third thing, especially in the world of-- think of compliance. For purpose of compliance, you would need individuals sitting there.
And really, more likely or not, a good nirvana state of the world would be where experts are working with the machines or experts are working with the AI rather than out of the loop. It won't be open loop system. Let me ask a question for the purpose of conversation. When was the first fully, self-driving, transcontinental flight take place? I think the first-- and I only know this-- I mean, this is probably not what you're asking, but I'll go for it because I felt like I sound smart.
The first transcontinental flight? Oh, shoot, I was going to say the first ever flight because I was just in Pennsylvania at a flight field where in 1896 was the first unmanned flight. Yes. But I think it only went 1,000 feet, so not transcontinental. No, no. So there's a plane that took off a runway, flew, and then landed. I don't know, in the last 30 years? Again, [INAUDIBLE] likely you know this. So it was late 1940s.
And so this is better and better technology and all that stuff, but still we always have the pilots there. And as we know, I mean, actually, people who do fly, as I understood from people who fly, flying is seriously difficult taxing business because you're taking your device into three dimensions rather than two dimension. And so maybe when you're flying in the air, when machines-- I mean, automation can take care of it by tracking things and all that stuff. Why tax humans?
But when landing and takeoff happens, which is actually highly uncontrolled setting, let's tax humans. So in a similar fashion, I think that might be the steady state we might be evolving towards continuously, right, in all aspects. And that's a good thing. And as we do those things, I think there will be more and more human jobs would be created. Expert jobs will be created around that. It's like there were no consultants maybe 40 or 50 years ago.
Now there are lots of consultants because there was this whole thing became a knowledge information economy. And to solve that, just because we had a lot of knowledge, didn't take people out. So I believe that in a similar fashion, there will be AI jobs that will be created where experts will have a role to play. And we'll need experts, not just to build more AI solutions, but also to work with AI solutions.
To remove people's fear of this-- because I think you got it right when you said at the beginning that fear is really what overpowers most of the general public when it comes to the concept of AI, and especially in the most recent months-- what would be a concrete example? You gave the example of people are washing clothes and then we have washing machines, and you need people to work the washing machines.
What would be some of the most concrete examples of where you see in the next 5 to 10 years of this transfer of jobs to AI where you still need people? So if I knew precise answer, Liberty, I will not tell you. I'll tell you why. Because then that's where I will take all my money and invest, OK. But I'll give you an approximate answer.
Pre-internet, if you had asked any one of us, would you have imagined that we'll be watching movies not in disks, or cassettes, or in cinema, we will be getting food delivered, clothes delivered, and furniture delivered, none of us would have believed that. But connectivity would bring certain speed, certain removal of barriers, and certain ability to connect to disconnected parties so that certain types of activities would become possible.
In a similar manner, in a simplest form of AI, what's going to happen is that there's a lot of information, and machines will help us sort through that information better. Machines will provide us better recommendations. And question is that would these recommendations create more jobs of some form? I do believe they will. I don't know exactly what forms.
So continuing the theme of fear, I think it's interesting the development that happened with large language models and Bard and ChatGPT and so forth. It challenges our distinctive intelligence language. That's what made us different from all the other species. We've got language. And language is expressive and so forth. And now ChatGPT can chat, and it can talk like a person. So some of that fear maybe is real. Some of that fear maybe is anticipated and so forth.
But definitely, we're encroaching a space now where we're a lot more afraid than we were November 2022. And so the question is, what are your thoughts? What should we be afraid of? So I think what we should be truly afraid of and maybe going back and picking up on that misinformation thread, right, that is where personas can be assumed and our inability to actually tell apart, I think, that's what I would be really, really afraid of here.
I mean, we, in the United States, we have election coming up. I'm actually really scared what we're going to see because if we have seen something with Facebook-related activities for the past decade, I feel it almost feels like with these new abilities, it might be just a small tip of a massive, massive iceberg. And that is something that I am truly afraid of. I'm not afraid of jobs. I'm not afraid of job creation, the value it will bring.
And maybe if we translate it back, I think it's about us thinking through carefully what does regulation mean. And in fact, if I just follow maybe also thoughts about privacy because in the same vein that you described, now that these complex language models, the more data they have on you, the more they can appear as a trusted agent and elicit even more data from you, which then becomes a vicious cycle in terms of the misinformation and the impact they can have on you. Absolutely.
And actually, this reminds me of, Munther, one of the other related project that you and I had that we started off with our visit to-- what is it? In New York. Right. Data markets. Yeah. And the whole thing is of right now I'm afraid of asking questions to ChatGPT because who knows? Maybe I'm revealing information that it will learn and then it will sell it to somebody else.
When people used to write exciting blogs online, the reason they wrote is to tell to the world that they're smart, B, create their brand, and C, not to sell it. Now ChatGPT and things like that have gone around, eaten up all of that, and then now they're pretending them. And that is an issue, which means it's tomorrow, I don't want to ask any question to ChatGPT that reveals any intelligent information.
And if it does, maybe I want it to pay to me, if somebody else is getting paying to ChatGPT because when Google search were free, OK, that's fine. I'm still OK. But the moment ChatGPT starts charging people for whatever the Premium package is, I want a 100th of the cent for the queries that serves my ends using my data. That's a very good point. I think that brings us to the very interesting question of lawmakers and heads of companies were discussing in DC what regulation should look like.
And with ChatGPT, I mean, I remember when it first came out, there was a big story about how they typed in, write a poem about Donald Trump, and it refused to write a poem about Donald Trump. But then it was write a poem about Joe Biden, and it wrote this glowing, wonderful poem about Joe Biden. Whatever anyone's politics are, it obviously has its own biases. Yes. And so what should the regulation look like? Should it be the government regulating?
Should there be some organization that comes in to regulate? What should that look like for AI? It's a very difficult question. But bottom line is that we need regulation. What we do not need-- and maybe I can start by what we do not need-- we definitely do not need a heavy-handed industry driving the decision.
For example, even in the simplest form of this was when internet became internet, the net neutrality has been a challenge where the massive internet service providers deciding who to charge how much. Well, I mean, if you're a big company, sure will survive. But then the moment you start charging me for my thing or stop serving my content as a small companies, we cannot grow. So we have to avoid that kind of issues for sure. The example you pointed out, that is appropriate freedom is needed.
But at the same time, the thing that cannot happen is wrong information being created. So maybe somehow information authentication is needed. And then going from there, nobody can stop research. You can't-- this whole viewing that AI is like nuclear weapons and hence nuclear non-proliferation treaty, we need AI non-proliferation treaty, that doesn't sound right to me at all. Maybe a controlled use of that might be useful, but again, that cannot be like that.
Just the way utilities are governed, maybe we should be governing some of these things, too. I mean, we still have not governed social media yet. I mean, the pact that we signed when we did social media was you give me your data, and you get the utility for free. Cell phones don't do that. They charge as money. And then we have a contract in the return saying that you are not going to look into my data. So I think some of those things need to thought out.
But I think we as academics have a much, much bigger responsibility here. In a sense, I feel that if some of my colleagues say that of all the interesting AI research is going to happen in industry, I think is completely wrong. But everything related to this has to happen in academia because academia is where we will have unbiased view. And for that reason, I'm actually a lot more optimistic about academia as well doing very well in terms of thinking about AI research than not.
It's not just about complicated neural network and yet another AI system. So in fact, I'm going to take this another level. And your speculation will be interesting because I think a lot of us potentially are still struggling with this. So regulation has to happen at every level of life. I mean, every technology needs to be regulated because technology can have a side that is detrimental. Any technology can do that. A car can become a bomb, and it can kill people and so forth.
So we need regulation. But we have a problem here of defining who owns what. I mean, information flow and a certain level of information flow is actually needed. But then at some point, when is my data actually belongs to me? When the electric company is measuring my consumption of electricity, do I own that data or the electric company owns that data? When I use Google for free, does Google own the data in exchange for the free service that they gave me? We have a problem of definitions.
And I think back to your point about the academic pursuit in this regulation, we don't have a framework. Yes. I don't understand the framework by which we're even discussing regulation. What is the framework? Do we have a sense of where that's heading? Fantastic question. So maybe let's reduce the scope so that we can have a conversation.
And reduction of scope would be in the context of, let's call it, recommendation systems because in my mind, recommendation systems are one of the earliest AI applications that has started interacting with society at large. You go to online marketplace like Amazon, and you buy things, and you are recommended certain things. And that's how actually Amazon is controlled whose products are sold or not, by the way, because it's a place where so many vendors come and sell.
That's how your price discrimination has taken place because different people are shown different prices at different times for the same product. And that's where your information about what you bought or not has created preferences that Amazon could have consumed and maybe sold to somebody else saying that, hey, person in this zip code, et cetera, has purchased these things, and let's sell it to somebody else. So there's a lot of that happening there.
Same is happening in the social media platforms and where information consumption and so on. Now, if we go back to that, let's say, social media platform because that's where misinformation has been at least believed to be rampant, there are few things happening. One is who is publishing information? Then two is what recommendation algorithms are utilized? And then the three is how the advertisements are deployed and who is subjected to. And so there are three parties here.
There's the people who are consumers, the platform who is, let's call it a social media platform, and then the content producer, whether it's an advertised content or whatnot. There's the third party. And these three things need to stick together. Platform has gotten all the powers right now. The contract has to implicitly happen between platform and consumers and explicitly in now somewhat monopolistic manner between platform and vendors or third party providers.
That's what is going on right now in today, for example, at Google. That is, did they have unfair advantage with the government claiming that they had unfair advantage in data, and that's how they could charge a lot to the advertisers. So maybe that's a three-party environment that shows up everywhere. In the platform, consumers and sellers are content providers. And maybe that's one good place to start in terms of thinking about regulation.
My knowledge is limited, but I'm pretty sure there is very good ways for two-party interaction and contracting that contract theory in economics, for example, [INAUDIBLE]. Law has certain ways of thinking about how do you define what are the boundaries of these kind of interactions. But I think with data and AI, I think there's a new dimension. So it's like people like us along working with them are learning a lot of their stuff and then bringing that together might be the right thing to do.
It's one thing to talk about the laws within the US that are able to govern this kind of stuff, but we obviously have a global work on AI and a global race, if you might put it, to AI. And even something like social media, TikTok has gotten banned in many places in the United States because of the data transfer to China. And so when you have these arguments for regulating AI within the US space, that's obviously a very different discussion than regulating it in the global space.
And the argument being that if we regulate it here in the US, somebody else is going to do it. China is going to leapfrog ahead of us, and we're going to be in terrible trouble. So how do you balance that? And how do you see this globally working in terms of regulation? Maybe two parts, one is development and progress in R&D, let's call it, in AI is not necessarily what we should be regulating.
Maybe I'm just naive in addition to being optimistic is that development of the technology is not destructive. It's a deployment of technology without regulation is destructive. So maybe we should worry about the deployment, not necessarily development. Deployment being destructive is both destructive to outside world and to inside world. And one of the reasons United States would like, for example, to do that is to retain sanity of the society within itself.
So tomorrow, if I'm a country which is outside, and I see that, yes, lack of regulation is going to remove the sanity of the society, I think maybe people would do that, too. So a lot of this we're taking from the perspective of some evilness happening and some country taking over and so forth. But we really are facing a problem of compliance and the law. A company says, my product does this.
Who's going to verify given the fact that the product in itself is learning and changing and so forth, and now we have this legal aspect of when it doesn't do the right thing, who's responsible, and who's verifying? Is this in the realm of regulation or maybe a question of law, a question of legal action? It's a new world. Yeah, no, I completely agree.
I think that's another reason why I'm optimistic about future of academia being very bright for a while because these questions would need a very fundamental take on them and also interdisciplinary take on them. Wouldn't it be really good and cool, for example, as internet came around, secure transfer became an issue? And that's [INAUDIBLE] who is going to certify it? Well, very soon and things like that, that came around and certified it.
And then there's a whole ecosystem has been built around it. And it has been a global ecosystem because just the way people in the United States don't want to lose their dollars when they're transferred, people in China and people in India and people in Switzerland, everywhere, nobody wants to lose that. So I feel that style of ecosystem will evolve because it will be about self-interest. Alignment would be necessary for self-interest preservation rather than not.
When you say the future of academics is bright in this sense and that experts are really going to be working with these AI systems, and it's not going to be this loop, what is your advice for young data scientists or people who may be looking to the future of what kind of jobs they could have that would work well in the AI space?
Do they need to go to places like MIT in order to be experts in this field, or are we going to start seeing technical institutes like we would if you want to be a welder, or a plumber, or you're going to be a data scientist? Where do you see that future going for academia in general and the public for the kind of jobs that they plan to have?
Fantastic question. So I think I'm going to hijack on one of the things I learned from Munther, and I thought it's a fantastic point is that so far we in the United States have been universities of higher education. Where I grew up in India and for example, where Munther grew up, there were always these technical institutions, as you said, variety of welders and all that. There are such institutions out there in the United States as well, but they're in a lot more scarcity.
And I think we just need to have that enabled. Now, one potential model is that there is no reason why MIT can't do that as well. Now, the question is that how does it do that well while preserving its core on one hand and have its ability to help support that at the same time? And that's something that we all have to figure out.
But again, I think core to your question was that education in AI is needed both as a user, as somebody to understand, as a broadly society, and as an opportunity for younger folks as well. So I think there's definitely a role for all of us to play there. [MUSIC PLAYING] Thank you for listening to this month's episode of Data Nation. You can get more information and listen to previous episodes at our website idss.mit.edu, or follow us on Twitter and Instagram @MITIDSS.
If you liked this podcast, please don't forget to leave us a review on Spotify, Apple, or wherever you get your podcasts. Thank you for listening to Data Nation from the MIT Institute of Data Systems and Society.