Qlik Connect 2024 | Qlik AI Council

theCUBE.net

Clips
News
More from Qlik Connect 2024

Nina Schick

Founder & CEO, Qlik AI Council Member

Tamang Ventures Limited

Dr. Rumman Chowdhury

CEO, Co-Founder & Qlik AI Council Member

Humane Intelligence

Kelly Forbes

Executive Director & Qlik AI Council Member

AI Asia Pacific Institute

Dr. Michael Bronstein

DeppMind Professor of AI & Qlik AI Council Member

University of Oxford

Qlik Council looks to accelerate responsible adoption of enterprise AI

In January, QlikTech International AB announced the formation of its inaugural AI Council, a set of four advisors with a mission to accelerate the responsible development of the Qlik AI-driven product portfolio.The council’s members were drawn from a community of AI experts and empowered to guide the company’s R&D direction and foster ethical and responsible use of Qlik’s AI.Members of the Qlik AI Council talk about their mission during theCUBE’s coverage of Qlik Connect

play_circle_outline The Evolving Role of AI in Economy: Necessity of Governance, Safety, and Responsible Testing

play_circle_outline The role of user experience and the potential for users to become data sources

play_circle_outline The potential for AI to be deployed in sectors that address important global issues like climate change

play_circle_outline Structured public feedback and the right to repair AI systems

Info

Qlik AI Council

Nina Schick

Founder & CEO, Qlik AI Council Member Tamang Ventures Limited

Dr. Rumman Chowdhury

CEO, Co-Founder & Qlik AI Council Member Humane Intelligence

Kelly Forbes

Executive Director & Qlik AI Council Member AI Asia Pacific Institute

Dr. Michael Bronstein

DeppMind Professor of AI & Qlik AI Council Member University of Oxford

search

>> Welcome back, everyone, to theCube's live performance here in Orlando, Florida. I'm John Furrier, host of theCube. Got the whole CUBE team here. We are here for Qlik Connect 2024. We've got a great panel session here with the AI Council newly formed this year. It's a team of experts who will be advising us here at the event on the future of AI. Thanks for coming on theCube. Really appreciate you guys. Thanks for coming on. Michael, former Twitter. You've been in deep mind on the research side. Congratulations, Nina, you were on stage. Good to see you. Thanks for coming on. So first question, what does the AI Council do? What is the mission of the AI Council? Kelly, we'll start with you.

>> So I think the idea is that it evolves as AI governance and all these questions around data. What are the best approaches, continue to evolve, right? So the idea is that we can provide guidance from different angles, right? Because I think we all bring different expertise and different ideas that we can share.

>> What's the philosophy around innovation and balancing the innovation with, I won't say scary concerns, but people are like, "Okay, where does this go? Is it safe? Is it explainable?" Let's get into that conversation because there's a lot of innovation yet to come. What do we say about this?

>> Well, Michael and I were just discussing before we went live, he being on the frontier of the actual research. The things that we're seeing now 10 years ago for AI researchers building these models, would've simply been inconceivable. So I think that there is certainly an understanding that what's happening now is moving at an exponential rate. And of course, each of us on the council has a very different view, but my view is certainly that we are seeing a changing of economic models, a changing of the way that we work. And this is largely to do with exponential technologies, AI at the heart. Now of course, that has so many opportunities, but at the same time, it's extremely disruptive and there are a lot of risks. And by the way, one of the risks I think, is not adopting AI, but you have to do so safely.

>> , you're a CUBE alumni, you've been on before. What's your focus? What's your position on the council? What's your focus these days?

>> Well, I think a lot about governance of AI systems, responsible use of data, but also how we can create new and innovative ways to test and create assurance for these modern and newer AI models. Now, generative AI models are quite different from your traditional machine learning AI models. And specifically from, if you think about a compliance perspective, we have a lot of new laws passing that mandate responsible use. But to be perfectly frank, the other part of the science is even figuring out how to test and evaluate them. So Nina's, totally correct in that we have to think about responsible governance, but I'll take it a step further. So I always use the phrase, brakes help you drive faster. And it seems paradoxical, but innovation and responsibility are not opposites of each other. You drive fast in the freeway because you know that if there is an accident, there is something concerning, you can stop your car. Now, good governance is exactly that. You cannot innovate without good governance.

>> I love that phrase.

>> And the infrastructure's changing too, so not always it's a straight highway or freeway.

>> It is not.

>> Sometimes you've got some corners to take. And so guardrails is a term, people kick that around, but we have a whole infrastructure changeover where you're seeing neural networks, you're seeing knowledge graphs, data, and how data's managed, distributed, programmed on. We hear on stage today, data supply chain. First conference I've ever been to where data supply chain was on a mainstream keynote. Okay, that implies some sort of explainability software supply chains use all the time in the security context. Are we getting in an era where we don't yet know what we don't know? What is the future of this infrastructure? What happens next?

>> I think you almost have to start building things for the world of AI. I think our former legal systems, ethical ideas might not be completely... We might have to adapt things as well, but we might need to create things, right? So, I'm part of a very exciting initiative happening in Dubai right now where the first regulation is coming in to look at data from that perspective. How do we now make sure that data flows, but also that is aligned with the AI and the new risks and potential opportunities as well? Right? Because we cannot regulate an almost from the same perspective that we used to form the technologies like 30, 60, a hundred years ago, right?

>> What's interesting about what you're saying is that it's like the old expression, don't play chess if you don't know what checkmate looks like. So, do we even know what's going to happen next? So what do you guys see as the key optimization now for the world to focus on? Just get building, let chaos reign, then reign in the chaos? What happens? What is the best approach to take? Because I think people are really excited. At the same time, the other side of the coin, people are like, "Whoa, slow down." Especially when you start getting into sovereign cloud and globalization. Data is now intellectual property. It could be chaos, but is that a good thing? Open question.

>> So probably getting rid of the hype, so focusing on the right problems. Probably less of the risks that maybe sound very, I don't know, science fiction, and focusing more on the real problems that AI can bring.

>> What are some of those problems?

>> So in generative AI, for example, this could be related to misinformation, so-called fake news. So when you cannot distinguish between content that is real and content that is generated. So different regulation or different technologies that allow you to tell the two apart. So that would be for example, one problem. Rather than existential risks and thinking of AI will become conscious and will kill us all in some way.

>> Well, the current form of AI today is going to increase more generic vanilla content, fake content. So, the ability to weaponize AI in that context is definitely a real problem. I think we're starting to see it today.

>> Exactly.

>> Absolutely. I read a book about it a few years ago, but increasingly we'll have synthetic content that's inundating the information ecosystem. And of course, it isn't that all synthetic content is malicious. But it's more that the integrity of the information ecosystem and the integrity of our public and electoral processes are already being coming under sustained assaults. And if one doesn't trust information and one doesn't trust processes, well that's tricky, both for businesses as well as society. This is going to be a prevalent theme in this year, a year of multiple elections and going forward. So I think authenticating information, authenticating content, and continuing to instill processes where you can be transparent and maintain trust with your customers is absolutely essential.

>> I think trust is a very key word here because that's really going to be the... If you take the current trend and extend it out, more fake news, more synthetic data, more stuff, the user experience is also going to change for the better. Productivity's going to increase, but also the expectations will change. So one of the things we always talk about on theCUBE and we riff on and we theorize, is, there'll be a flight to quality and that users will start being involved. And so, I'm intrigued by the whole GameStop stock thing because a Reddit kind of like game culture. That the collective intelligence now of people, the user interactions as data, is that a direction that you guys see, where the user experience itself becomes a data source and how we consume content like a live stream-

>> Isn't it already?

>> Right .

>> It is already.

>> It already is.

>> Reddit is a primary source of information for generative AI models. For better or for worse, our behaviors, our actions, our attitudes on the internet are feeding and have always fed. I mean, even if we think of the concept of quote, "data exhaust", which is from the early two thousands, it was called exhaust because it was considered to be the useless information. But there it was. And this is where, as we think of quality of data, to your point, generative AI makes it easier to use data. Now, what will be valuable is the quality, the authenticity of the data, knowing that it is actually factual or comes from a human being, that the premium of that's going to be even higher.

>> And that's the importance of the breaks, which I really like that analogy. Because people for a very long time, think innovation versus regulation. And it's not necessarily right. Any strong governance or strong breaks can actually help that to make it sustainable.

>> Absolutely.

>> We have to start an AI safety board. If we're going to have highways and freeways and guardrails, might as well have some sort of safety system. But no, safety's being discussed a lot. And again, I bring that up tongue in cheek, but the reality is that the world's changing. And you brought up the fake news and people's role. Are we just one big node in the network? Is it very black mirror to say that there's a world where we have data. And if you look at Facebook, what they've done, they've taken users data and they've monetized the hell out of it. I own my own data. Why can't I be a node in the graph? A graph is just nodes in arcs, connecting us all together. Here on theCUBE, you're now CUBE alumni. We're in theCUBE network.

>> I was recently at an event with Charlie Brooker who obviously wrote Black Mirror. And his view was that when they came up with the episodes, they couldn't imagine this would all mirror so quickly in reality. But I think the point that you raised is really valid in the sense that all of these questions, as we enter this age of exponential disruption where technology, which has already so profoundly changed and impacted our lives and economic realities over the past 30 years in particular, is going to touch the lives of billions of people in an extremely personal way. And it's going to become a highly political and politicized issue. So this isn't only about business transformation, but societal transformation. And the issue that you raise about one's own digital sovereignty, I think, is going to become a very important political issue.

>> Well, I also wonder how much of that is education? Would people even know what to do if they woke up and Google said, "Actually, you own all of your data." So we've already had a mini-experiment with it, with people trying to push decentralized social media. First of all, people didn't know what it meant. Second, they thought it was a high friction, high barrier to entry. And they got frustrated and confused and kind of went to their usual centralized models. Now, the purpose of decentralization was that, was to give people more ownership of their experience, what they're doing. And this is not to say that it's not a good idea, it's that we are not accustomed as a society to having that level of ownership. Now, one of my favorite books in this whole field is actually Shoshana Zuboff's, the Age of Surveillance Capitalism.

>> I love that.

>> It's a great book because it's about the economics of tech, and that's what she talks about. It is designed to, it's a data consumption economy. So if we're truly going to understand data sovereignty and ownership, we have to figure out how to do that in a way that understands the economic and societal constraints as well as the consumer knowledge and understanding.

>> That's a great point about people not knowing what to do with the data. It points to the user experience expectations challenge. With every new wave, there's always new expectations. I mean mobile devices and then iPhone, and then now we got this. So the question is, what's the benefit to the user? At the end of the day, you know what? All right, Facebook, all my photos are there. All right, whatever. It's centralized, but am I really going to... And some people do quit Facebook. But the point is it's easier for me to keep it centralized, to your point. What has to change? I guess the question is, what does the human society do? Because in the loop is key to the data. As an AI council, if you guys could have your wish, what should people do to engage as a society or with the council to help be more educated, to help have a conversation? Because I think it's a democratic view here, that if everyone's looking at it, let chaos reign. And then reign in the chaos with the guardrails and the view. What should people do? How do people get involved? What would you guys recommend?

>> So if I can say, so I don't think that we're talking about some phenomenon that is happening. So this is technology that we are developing, so we need to control it and decide. So certain developments are choices and the way that this technology is used by the human society. So my concern if we're talking about somewhat slightly apocalyptic scenarios is, if we get used to AI, especially in creative intellectual tasks like creative writing or maybe artistic work, that we'll lose our skills. And we have, over the course of history or evolution of civilization, certain skills that we've lost. Like, who nowadays can produce fire? But this is probably the first time that we are talking about skills that are considered the hallmark of humans. So intellectual, creative skills. And to me, this is scary, so it's not a kind of extinction scenario. But do we want to live in a world where all the cool stuff is done by machines and we're left to do what?

>> And just to add to that, I feel that-

>> That's a good point....

>> our education systems and all systems around us, are not necessarily adapting as fast as the technologies to this change. So if you think about our kids at school, they're still learning in the same way that they used to a hundred years ago. So how do we adapt and change that, prepare people for this revolution that we are about to embark on? Right?

>> Yeah. I mean the killer app is creativity and time freeing up to do cool things.

>> That's right.

>> Right. So, good point about doing cool things. We don't want machines to do the cool stuff. The boring stuff, maybe, but that's a great point. So, final question for each of you. If you could connect the dots next couple years, what would be the ideal scenario for AI? Kelly, we'll start with you.

>> Wow. I think we would like to see AI being deployed in sectors that really matter, and where we really need to see. Climate change, we look at sustainable goals where we have very little time to accomplish those goals. That's where AI comes in, because AI has the potential to accelerate that, make it possible. To do that, I think that we're still living in a time where there's a lot of hesitation because of the risks. Because of every day you see on the headlines there's something going wrong. So, that's where I think guard rails and thinking about, what are the right breaks that we need to have in place so that people feel confident enough to really use and leverage this technology?

>> Check your breaks. Good point. ?

>> The big thing I focus on is something I call structured public feedback. How can feedback from regular people... If we're talking general purpose AI models, they have output perspectives that may impact every person, every aspect of their lives. So with generative AI, we actually have an increased ability to engage with people as part of the no-code revolution. So I'd want to see, is something I'm advocating for, an AI right to repair. What does it mean for you and me and Kelly and Nina and Michael to individually repair and fix our own AI systems? And how can we make that something accessible to everybody on earth?

>> That's a noble mission.

>> So what I conceive of what's happening is almost like a new industrial revolution, where we're going to be producing intelligence. You talked about the infrastructure, both digital and physical. That's already being built, that's already being integrated into every facet of the lives of billions of people around the world. So, what I'm really interested in ultimately comes down to power and accountability. Who are the players or who are the powers that own this infrastructure? And who are they accountable to? Rather than thinking about AI as autonomously taking over the world-

>> ....

>> or changing our lives-

>> ....

>> I think it's very much still driven by human interests and economic interests. So, it's understanding how that impacts all of us.

>> Awesome. Well said.

>> So for me, well in addition to industrial revolution I would like, I'm excited, but at the same time, scared about scientific revolution. So the scientific method that we currently use, and it has taken us, I think to the place where we are, I think quite incredible over what, the course of probably four centuries. We are probably for the first time, challenging this view. So AI for science or AI-based or AI-driven science is very different from the way that science has been done in the past. And we also see it in the way that experiments are done, the data is collected. So data is collected probably for the first time systematically to be used by AI, not by humans. So, it can probably lead us to faster and bigger breakthroughs. But it is not clear what will it make to humans? So will there be still need for human scientists, for example.

>> Well, I appreciate the council and you guys coming on sharing your expertise with us. Thank you very much. And I'll go a little bit over segment here. For the last round, take a minute or 30 seconds or so to put a plug in for what you're excited about working on now. What's your current focus, what you're excited about? Give a quick plug for your work. Kelly, we'll start with you.

>> I'm very excited to see how developing countries can leverage this technology for good, how they can be deployed in sectors that we see flourishing. And the potential that we have there to really think of a future where AI is being deployed for good. Right?

>> So, I run a non-profit called Humane Intelligence. We're building an open source test and evaluation platform. We're also trying to educate and build a curriculum for the next generation of algorithmic assessors and auditors. And as Michael pointed out, there's a science to evaluations that we're uncovering. So figuring out best practices and scientific validity around red teaming, evals, bias bounties. This is critical for good governance. We have to get the science first.

>> I work with some of the most interesting frontier AI companies in the world, and leaders looking to understand how to interpret what's happening now in this era of exponential tech-led change. I think that this epoch is the most interesting epoch in human history. So the question is really, how are we going to deal with this huge transformation, this huge disruption that is upon us in society? I think that's the challenge for each of us, because I think our lives are going to be very, very different in five years, 10 years, 50 years from now.

>> So, I'm a computer scientist. We are developing new generative AI technologies, in particular for applications in life science and biology applications like discovery and development of new drugs.

>> Congratulations. Thank you so much. I learned a lot. I feel smarter already. Thank you for coming on.

>> Thanks for having us.

>> It's a wild west ahead, as they say, but I think it's an exciting time. Keep the brakes on, go faster. Got to get those algorithms in check, and I love that mission, and I love what you guys are doing. Thank you for coming on theCUBE and sharing your work and your commentary with our community.

>> I appreciate you.

>> Okay. Live here in Orlando, I'm John Furrier with theCUBE. Thanks for watching. We'll be back with more coverage after this short break.

Qlik AI Council

search

>> Welcome back, everyone, to theCube's live performance here in Orlando, Florida. I'm John Furrier, host of theCube. Got the whole CUBE team here. We are here for Qlik Connect 2024. We've got a great panel session here with the AI Council newly formed this year. It's a team of experts who will be advising us here at the event on the future of AI. Thanks for coming on theCube. Really appreciate you guys. Thanks for coming on. Michael, former Twitter. You've been in deep mind on the research side. Congratulations, Nina, you were on stage. Good to see you. Thanks for coming on. So first question, what does the AI Council do? What is the mission of the AI Council? Kelly, we'll start with you.

>> So I think the idea is that it evolves as AI governance and all these questions around data. What are the best approaches, continue to evolve, right? So the idea is that we can provide guidance from different angles, right? Because I think we all bring different expertise and different ideas that we can share.

>> What's the philosophy around innovation and balancing the innovation with, I won't say scary concerns, but people are like, "Okay, where does this go? Is it safe? Is it explainable?" Let's get into that conversation because there's a lot of innovation yet to come. What do we say about this?

>> Well, Michael and I were just discussing before we went live, he being on the frontier of the actual research. The things that we're seeing now 10 years ago for AI researchers building these models, would've simply been inconceivable. So I think that there is certainly an understanding that what's happening now is moving at an exponential rate. And of course, each of us on the council has a very different view, but my view is certainly that we are seeing a changing of economic models, a changing of the way that we work. And this is largely to do with exponential technologies, AI at the heart. Now of course, that has so many opportunities, but at the same time, it's extremely disruptive and there are a lot of risks. And by the way, one of the risks I think, is not adopting AI, but you have to do so safely.

>> , you're a CUBE alumni, you've been on before. What's your focus? What's your position on the council? What's your focus these days?

>> Well, I think a lot about governance of AI systems, responsible use of data, but also how we can create new and innovative ways to test and create assurance for these modern and newer AI models. Now, generative AI models are quite different from your traditional machine learning AI models. And specifically from, if you think about a compliance perspective, we have a lot of new laws passing that mandate responsible use. But to be perfectly frank, the other part of the science is even figuring out how to test and evaluate them. So Nina's, totally correct in that we have to think about responsible governance, but I'll take it a step further. So I always use the phrase, brakes help you drive faster. And it seems paradoxical, but innovation and responsibility are not opposites of each other. You drive fast in the freeway because you know that if there is an accident, there is something concerning, you can stop your car. Now, good governance is exactly that. You cannot innovate without good governance.

>> I love that phrase.

>> And the infrastructure's changing too, so not always it's a straight highway or freeway.

>> It is not.

>> Sometimes you've got some corners to take. And so guardrails is a term, people kick that around, but we have a whole infrastructure changeover where you're seeing neural networks, you're seeing knowledge graphs, data, and how data's managed, distributed, programmed on. We hear on stage today, data supply chain. First conference I've ever been to where data supply chain was on a mainstream keynote. Okay, that implies some sort of explainability software supply chains use all the time in the security context. Are we getting in an era where we don't yet know what we don't know? What is the future of this infrastructure? What happens next?

>> I think you almost have to start building things for the world of AI. I think our former legal systems, ethical ideas might not be completely... We might have to adapt things as well, but we might need to create things, right? So, I'm part of a very exciting initiative happening in Dubai right now where the first regulation is coming in to look at data from that perspective. How do we now make sure that data flows, but also that is aligned with the AI and the new risks and potential opportunities as well? Right? Because we cannot regulate an almost from the same perspective that we used to form the technologies like 30, 60, a hundred years ago, right?

>> What's interesting about what you're saying is that it's like the old expression, don't play chess if you don't know what checkmate looks like. So, do we even know what's going to happen next? So what do you guys see as the key optimization now for the world to focus on? Just get building, let chaos reign, then reign in the chaos? What happens? What is the best approach to take? Because I think people are really excited. At the same time, the other side of the coin, people are like, "Whoa, slow down." Especially when you start getting into sovereign cloud and globalization. Data is now intellectual property. It could be chaos, but is that a good thing? Open question.

>> So probably getting rid of the hype, so focusing on the right problems. Probably less of the risks that maybe sound very, I don't know, science fiction, and focusing more on the real problems that AI can bring.

>> What are some of those problems?

>> So in generative AI, for example, this could be related to misinformation, so-called fake news. So when you cannot distinguish between content that is real and content that is generated. So different regulation or different technologies that allow you to tell the two apart. So that would be for example, one problem. Rather than existential risks and thinking of AI will become conscious and will kill us all in some way.

>> Well, the current form of AI today is going to increase more generic vanilla content, fake content. So, the ability to weaponize AI in that context is definitely a real problem. I think we're starting to see it today.

>> Exactly.

>> Absolutely. I read a book about it a few years ago, but increasingly we'll have synthetic content that's inundating the information ecosystem. And of course, it isn't that all synthetic content is malicious. But it's more that the integrity of the information ecosystem and the integrity of our public and electoral processes are already being coming under sustained assaults. And if one doesn't trust information and one doesn't trust processes, well that's tricky, both for businesses as well as society. This is going to be a prevalent theme in this year, a year of multiple elections and going forward. So I think authenticating information, authenticating content, and continuing to instill processes where you can be transparent and maintain trust with your customers is absolutely essential.

>> I think trust is a very key word here because that's really going to be the... If you take the current trend and extend it out, more fake news, more synthetic data, more stuff, the user experience is also going to change for the better. Productivity's going to increase, but also the expectations will change. So one of the things we always talk about on theCUBE and we riff on and we theorize, is, there'll be a flight to quality and that users will start being involved. And so, I'm intrigued by the whole GameStop stock thing because a Reddit kind of like game culture. That the collective intelligence now of people, the user interactions as data, is that a direction that you guys see, where the user experience itself becomes a data source and how we consume content like a live stream-

>> Isn't it already?

>> Right .

>> It is already.

>> It already is.

>> Reddit is a primary source of information for generative AI models. For better or for worse, our behaviors, our actions, our attitudes on the internet are feeding and have always fed. I mean, even if we think of the concept of quote, "data exhaust", which is from the early two thousands, it was called exhaust because it was considered to be the useless information. But there it was. And this is where, as we think of quality of data, to your point, generative AI makes it easier to use data. Now, what will be valuable is the quality, the authenticity of the data, knowing that it is actually factual or comes from a human being, that the premium of that's going to be even higher.

>> And that's the importance of the breaks, which I really like that analogy. Because people for a very long time, think innovation versus regulation. And it's not necessarily right. Any strong governance or strong breaks can actually help that to make it sustainable.

>> Absolutely.

>> We have to start an AI safety board. If we're going to have highways and freeways and guardrails, might as well have some sort of safety system. But no, safety's being discussed a lot. And again, I bring that up tongue in cheek, but the reality is that the world's changing. And you brought up the fake news and people's role. Are we just one big node in the network? Is it very black mirror to say that there's a world where we have data. And if you look at Facebook, what they've done, they've taken users data and they've monetized the hell out of it. I own my own data. Why can't I be a node in the graph? A graph is just nodes in arcs, connecting us all together. Here on theCUBE, you're now CUBE alumni. We're in theCUBE network.

>> I was recently at an event with Charlie Brooker who obviously wrote Black Mirror. And his view was that when they came up with the episodes, they couldn't imagine this would all mirror so quickly in reality. But I think the point that you raised is really valid in the sense that all of these questions, as we enter this age of exponential disruption where technology, which has already so profoundly changed and impacted our lives and economic realities over the past 30 years in particular, is going to touch the lives of billions of people in an extremely personal way. And it's going to become a highly political and politicized issue. So this isn't only about business transformation, but societal transformation. And the issue that you raise about one's own digital sovereignty, I think, is going to become a very important political issue.

>> Well, I also wonder how much of that is education? Would people even know what to do if they woke up and Google said, "Actually, you own all of your data." So we've already had a mini-experiment with it, with people trying to push decentralized social media. First of all, people didn't know what it meant. Second, they thought it was a high friction, high barrier to entry. And they got frustrated and confused and kind of went to their usual centralized models. Now, the purpose of decentralization was that, was to give people more ownership of their experience, what they're doing. And this is not to say that it's not a good idea, it's that we are not accustomed as a society to having that level of ownership. Now, one of my favorite books in this whole field is actually Shoshana Zuboff's, the Age of Surveillance Capitalism.

>> I love that.

>> It's a great book because it's about the economics of tech, and that's what she talks about. It is designed to, it's a data consumption economy. So if we're truly going to understand data sovereignty and ownership, we have to figure out how to do that in a way that understands the economic and societal constraints as well as the consumer knowledge and understanding.

>> That's a great point about people not knowing what to do with the data. It points to the user experience expectations challenge. With every new wave, there's always new expectations. I mean mobile devices and then iPhone, and then now we got this. So the question is, what's the benefit to the user? At the end of the day, you know what? All right, Facebook, all my photos are there. All right, whatever. It's centralized, but am I really going to... And some people do quit Facebook. But the point is it's easier for me to keep it centralized, to your point. What has to change? I guess the question is, what does the human society do? Because in the loop is key to the data. As an AI council, if you guys could have your wish, what should people do to engage as a society or with the council to help be more educated, to help have a conversation? Because I think it's a democratic view here, that if everyone's looking at it, let chaos reign. And then reign in the chaos with the guardrails and the view. What should people do? How do people get involved? What would you guys recommend?

>> So if I can say, so I don't think that we're talking about some phenomenon that is happening. So this is technology that we are developing, so we need to control it and decide. So certain developments are choices and the way that this technology is used by the human society. So my concern if we're talking about somewhat slightly apocalyptic scenarios is, if we get used to AI, especially in creative intellectual tasks like creative writing or maybe artistic work, that we'll lose our skills. And we have, over the course of history or evolution of civilization, certain skills that we've lost. Like, who nowadays can produce fire? But this is probably the first time that we are talking about skills that are considered the hallmark of humans. So intellectual, creative skills. And to me, this is scary, so it's not a kind of extinction scenario. But do we want to live in a world where all the cool stuff is done by machines and we're left to do what?

>> And just to add to that, I feel that-

>> That's a good point....

>> our education systems and all systems around us, are not necessarily adapting as fast as the technologies to this change. So if you think about our kids at school, they're still learning in the same way that they used to a hundred years ago. So how do we adapt and change that, prepare people for this revolution that we are about to embark on? Right?

>> Yeah. I mean the killer app is creativity and time freeing up to do cool things.

>> That's right.

>> Right. So, good point about doing cool things. We don't want machines to do the cool stuff. The boring stuff, maybe, but that's a great point. So, final question for each of you. If you could connect the dots next couple years, what would be the ideal scenario for AI? Kelly, we'll start with you.

>> Wow. I think we would like to see AI being deployed in sectors that really matter, and where we really need to see. Climate change, we look at sustainable goals where we have very little time to accomplish those goals. That's where AI comes in, because AI has the potential to accelerate that, make it possible. To do that, I think that we're still living in a time where there's a lot of hesitation because of the risks. Because of every day you see on the headlines there's something going wrong. So, that's where I think guard rails and thinking about, what are the right breaks that we need to have in place so that people feel confident enough to really use and leverage this technology?

>> Check your breaks. Good point. ?

>> The big thing I focus on is something I call structured public feedback. How can feedback from regular people... If we're talking general purpose AI models, they have output perspectives that may impact every person, every aspect of their lives. So with generative AI, we actually have an increased ability to engage with people as part of the no-code revolution. So I'd want to see, is something I'm advocating for, an AI right to repair. What does it mean for you and me and Kelly and Nina and Michael to individually repair and fix our own AI systems? And how can we make that something accessible to everybody on earth?

>> That's a noble mission.

>> So what I conceive of what's happening is almost like a new industrial revolution, where we're going to be producing intelligence. You talked about the infrastructure, both digital and physical. That's already being built, that's already being integrated into every facet of the lives of billions of people around the world. So, what I'm really interested in ultimately comes down to power and accountability. Who are the players or who are the powers that own this infrastructure? And who are they accountable to? Rather than thinking about AI as autonomously taking over the world-

>> ....

>> or changing our lives-

>> ....

>> I think it's very much still driven by human interests and economic interests. So, it's understanding how that impacts all of us.

>> Awesome. Well said.

>> So for me, well in addition to industrial revolution I would like, I'm excited, but at the same time, scared about scientific revolution. So the scientific method that we currently use, and it has taken us, I think to the place where we are, I think quite incredible over what, the course of probably four centuries. We are probably for the first time, challenging this view. So AI for science or AI-based or AI-driven science is very different from the way that science has been done in the past. And we also see it in the way that experiments are done, the data is collected. So data is collected probably for the first time systematically to be used by AI, not by humans. So, it can probably lead us to faster and bigger breakthroughs. But it is not clear what will it make to humans? So will there be still need for human scientists, for example.

>> Well, I appreciate the council and you guys coming on sharing your expertise with us. Thank you very much. And I'll go a little bit over segment here. For the last round, take a minute or 30 seconds or so to put a plug in for what you're excited about working on now. What's your current focus, what you're excited about? Give a quick plug for your work. Kelly, we'll start with you.

>> I'm very excited to see how developing countries can leverage this technology for good, how they can be deployed in sectors that we see flourishing. And the potential that we have there to really think of a future where AI is being deployed for good. Right?

>> So, I run a non-profit called Humane Intelligence. We're building an open source test and evaluation platform. We're also trying to educate and build a curriculum for the next generation of algorithmic assessors and auditors. And as Michael pointed out, there's a science to evaluations that we're uncovering. So figuring out best practices and scientific validity around red teaming, evals, bias bounties. This is critical for good governance. We have to get the science first.

>> I work with some of the most interesting frontier AI companies in the world, and leaders looking to understand how to interpret what's happening now in this era of exponential tech-led change. I think that this epoch is the most interesting epoch in human history. So the question is really, how are we going to deal with this huge transformation, this huge disruption that is upon us in society? I think that's the challenge for each of us, because I think our lives are going to be very, very different in five years, 10 years, 50 years from now.

>> So, I'm a computer scientist. We are developing new generative AI technologies, in particular for applications in life science and biology applications like discovery and development of new drugs.

>> Congratulations. Thank you so much. I learned a lot. I feel smarter already. Thank you for coming on.

>> Thanks for having us.

>> It's a wild west ahead, as they say, but I think it's an exciting time. Keep the brakes on, go faster. Got to get those algorithms in check, and I love that mission, and I love what you guys are doing. Thank you for coming on theCUBE and sharing your work and your commentary with our community.

>> I appreciate you.

>> Okay. Live here in Orlando, I'm John Furrier with theCUBE. Thanks for watching. We'll be back with more coverage after this short break.