Hanh Brown: 00:00:00
Hi, I’m Hanh Brown, your host of the Boomer Living Broadcast. Well, I’m here to share an event that I took part in organized by Will Trevor from Tufts University’s Gordon Institute. The focus of our discussion was “AI: Is ChatGPT Overhyped and Underrated, or Underhyped and Underestimated?” Well, we explored ChatGPT, a notable AI platform from OpenAI.
Hanh Brown: 00:00:26
Since its launch in late 2022, chatGPT has spurred much discussion among both experts and hobbyists serving as a key point of reference in the considering the future of artificial intelligence and a societal impact. Well, in the debate we ask, is ChatGPT getting more attention and praise than it merits or is its potential underestimated? We had a panel of leaders in the field
Hanh Brown: 00:00:52
offering different perspectives on ChatGPT’s abilities and potential. Our goal was to provide a well-rounded exploration of this topic. The debate highlighted various viewpoints of ChatGPT, ranging from skeptics, questioning its capacity to emulate human intelligence to visionaries, imagining a future reshape by it. So join us as we delve into this topic.
Will Trevor: 00:01:20
So, hi everyone, and uh, and welcome. Uh, my name is, uh, Will Trevor. I’m Director of Programs and Operations, uh, here at, uh, Tufts Gordon Institute, uh, at uh, at Tufts University. Welcome to you. Uh, welcome to the today’s keynote debate. Uh, the topic is “AI: Is ChatGPT Overhyped and Overrated or
Will Trevor: 00:01:43
Underhyped and Underestimated?” Uh, now the aim of, uh, these debates is to ring together a group of, uh, leading thinkers, uh, both practitioners and academics, to discuss and debate a topical issue. At the intersection of leadership, uh, innovation management and entrepreneurship, our aim is to help nurture and develop the next generation
Will Trevor: 00:02:07
of transformative leaders, uh, with, uh, with pro, who act with profound purpose. And who are informed by the relevant topics of the day. Uh, first, however, I’d like to call upon Kevin Oye, uh, who is the Executive Director of Tufts Gordon Institute, uh, which is also home to the Darby Entrepreneurship Center and the Oster Center for
Will Trevor: 00:02:30
Applied uh, innovation Research. Uh, just call upon Kevin just to say a few words by way of introduction.
Kevin Oye: 00:02:36
Yeah. Thank you, Will. And thank you all to the panelists for joining today and to all the attendees. You know, we’re living in a very crazy time now where there’s so much going on in the world. You can take about world events, you can talk about technology trends, you can talk about social trends, but the richness of
Kevin Oye: 00:02:49
what’s going on is creating the demand for all of us, each of us, to become much, much more conversant in emerging technologies such as ChatGPT, or uh, AI algorithms, machine learning algorithms. Because in many ways technology touches everyone and everything in our everyday life. You know, there are more people in the world today with mobile telephones than
Kevin Oye: 00:03:13
there are people with indoor plumbing. You know, it only took ChatGPT three days to get to a million subscribers when I took Netflix over a year to get to that point. And all of us are interacting every day with AI ML machines because we interact with machines that serve up adds to us, serve up things that we should buy, serve up streaming video that we should look
Kevin Oye: 00:03:29
at based on what we looked at before. These are all applications of machine learning. And so the, the, the latest invention of that, of course is ChatGPT, which is basically a sentence completion machine. It’s a way of doing predictive response in terms of, uh, natural language processing. But the more important thing is by doing series like this, we wanna give
Kevin Oye: 00:03:50
everyone an opportunity to hear all facets, all aspects of a particular technology and its impact on society. Our goal is not to give you the answer or to, or, or advocate one way or the other. Our goal is to create an open forum. And so, I wanna thank all of you on the panel today for coming because all of you have very strong opinions and I hope you will share them.
Kevin Oye: 00:04:11
And then for everyone in the audience, I hope you’ll listen to all sides with an open mind and challenge our panelists too, with questions. Because, I think through that we’ll all become much more educated and we’ll all become much more wiser consumers and also wiser citizens of the world. Because as we make choices in the world about what things we wanna support, what
Kevin Oye: 00:04:30
things we wanna buy, what things we wanna regulate, through laws, et cetera, the more we each know about these areas, the better we can be at applying our own uprights and responsibilities as citizens. So, thank you very much for taking the time today. I look forward to the panel.
Will Trevor: 00:04:47
Thanks, Kevin. Uh, so here at the Tufts Gordon Institute (recording in progress), we offer a range of, uh, different programs from, uh, undergraduate through to professional education. Uh, recently we launched a redesigned MST in engineering management that leverages the best of, uh, our award-winning on campus and online
Will Trevor: 00:05:07
programs, uh, and re envisage it it into a single, top tier graduate program that prepares leaders for the new world of work and available 100% online. Uh, not only does the Tufts NSEM, uh, deliver the relevant content and practical skills you need the thrive in an increasingly complex and rapidly changing world, but it also provides you with an exceptional part-time graduate
Will Trevor: 00:05:31
experience that can fit into busy life. Now the LS in, uh, Innovation and Management. It’s a full-time, uh, accelerated program designed for people who want to make things happen, who are passionate about science, technology, engineering, or math, but who believe new solutions will solve our toughest challenges, uh, and make this a better world.
Will Trevor: 00:05:51
If you’re a recent STEM graduate, uh, who wants to accelerate your career and be at the forefront of developing new ideas, the MSIM, the MS in Innovation Management Program will provide you with the product innovation, uh, marketing, finance, and leadership skills. You need to be a new leader in today’s technology driven world. Um, just, uh, a few words on the
Will Trevor: 00:06:10
format of the event in, in a moment. Uh, I’ll call up on each of the panelists in turn to make about a three to five minute pitch, either four, uh, or, uh, against the motion under discussion. Uh, then around one o’clock Eastern Time. Uh, we’ll start a discussion among the panelists. Uh, we’d also like to include some questions from you, the audience.
Will Trevor: 00:06:33
So, uh, there is the q and a button there. Uh, if you’d like to post your questions there. Um, we may not be able to get to all of them, but we’ll certainly incorporate as many as, uh, as we can. Um, so, uh, by now I, firstly, I’d just like to, uh, introduce you to the panelists. I’m not gonna read their, all of
Will Trevor: 00:06:51
them have very full, uh, biographies. Uh, and my colleague Jill will be posting, uh, a pdf f with all their, their b biographies on there. But very briefly, um, uh, welcome Hanh Brown. Uh, now Hanh is a dedicated Age Technologist and AI Innovator, uh, and she’s dedicated to transforming the, uh, the aging, uh, experience.
Will Trevor: 00:07:14
Um, we have, uh, for the second time, he joined us for a very first, um, keynote debate for last year, uh, Bhaskar Chakravorti, uh, who is Dean of Global Business, uh, at the Fletcher School here at Tufts. And, uh, Founding Executive, uh, uh, Director of Fletcher’s Institute for Business in the global context. We’re also very pleased to be joined by,
Will Trevor: 00:07:34
uh, our very own Partha Ghosh, uh, member of faculty for Tufts Gordon Institute. Uh, he’s a well-renowned sport Thought Leader, uh, Philosopher, uh, Problem Solver, Futurist and Inspirer in the field of, uh, Business, uh, and, uh, and Technology. Uh, also joining us, a member of, uh, Tufts faculty, uh, James Intriligator. Uh, James is a Professor of the Practice,
Will Trevor: 00:07:57
uh, and Director of Strategic Innovation, uh, in the Department of Mechanical Engineering, uh, in the School of Engineering here at, uh, Tufts University. Uh, and also joining us, we’re very fortunate to have, uh, Lori Mazor. Now Lori uh, is an Architect and Visionary with a career in higher education spanning three decades. Uh, she has Planning, Design, and
Will Trevor: 00:08:17
Operations at, at some of New York City’s leading academic institute institutions, including New York University and City University of New York. Um, panelists, welcome to, uh, all of you, uh, panelists, if I can r remind you that we have an informed, uh, audience, um, but they’re not necessarily a, uh, a specialist, uh, audience in, uh, uh, uh, in AI and ChatGPT.
Will Trevor: 00:08:44
So, uh, if there is anything that you think is a. An injury industry term, or a concept that might be somewhat, uh, difficult to grasp if you’re not a specialist, please do, please do, uh, explain it. Um, uh, so first of all, uh, if I could call upon you Hanh uh, Brown, uh, if you could start us with your sort of circle three to five minute,
Will Trevor: 00:09:03
uh, uh, discussion on the topic?
Hanh Brown: 00:09:06
Sure. Thank you. Thank you so much, uh, for asking me to be here. Well, um, well, I think ChatGPT represents significant leap in artificial intelligence. Now, despite some current limitations, it underscores the technological advancements, particularly in the realm of
Hanh Brown: 00:09:21
language, machine language understanding. Well, imagine a world where machines not only understand our language, but also generate text similar to a human. So I think it’s, it’s an exciting prospect that is closer than what we might think. And thanks to the advancement of ChatGPT. So I also think that it offers a glimpse into the boundless potential of AI in the creative industry, which
Hanh Brown: 00:09:49
means the text generated by this model is very coherent, creative, and surprisingly very human-like. And it also demonstrate the vast capabilities of AI, but also it symbolizes this progressive pursuit of technological advancements. Now, from a business perspective, well, I think it holds the potential to disrupt traditional operations, let’s
Hanh Brown: 00:10:17
say applications and customer service. It can provide accurate, fast responses and offer support around the clock significantly enhancing customer experience. So that’s a lot of savings, especially nowadays that we’re shortage on staff. Now, it’s aptitude for generating text suggests transformative impact on content creation.
Hanh Brown: 00:10:46
So, for writers, they could leverage this technology to generate ideas, draft initial versions, or even produce ready to published articles. So I think it’s, it’s very exciting now. It also has a potential to democratize AI, making it accessible to a much wider audience. So, whether it’s automating tasks to increasing productivity, or facilitating
Hanh Brown: 00:11:15
more personalized learning experiences, I think ChatGPT is a beacon of the kind of societal adaptation that’s required as we progress technologically. So, as far as the future, I think ChatGPT, not, not as, um, the end product, but more like, um, Foundation Stone for more Advanced AI systems, yet to come. So it, it’s exciting and I also
Hanh Brown: 00:11:42
look forward to how the technology underpinning ChatGPT could be employed in other AI applications and paving the way even more innovative tools. So, I’m using it daily, learning, growing, innovating, and every day is a new day because I feel, whatever I thought I knew last week, it’s already has changed. So, I’m excited for this technology and much more to come.
Will Trevor: 00:12:16
Thanks for getting us started, Hanh. And uh, next up Bhaskhar, can I invite you to, uh, speak to the, uh, the topic?
Bhaskar Chakravorti: 00:12:27
Thank you, will thank you, uh, to you and all your colleagues for inviting us for this, uh, very timely conversation. And I’m hoping to learn, uh, a lot from my, uh, fellow panelists or fellow debaters here. And Hanh, thank you for, uh, uh, kicking us off. Uh, though I noted that you did not,
Bhaskar Chakravorti: 00:12:42
uh, vote on, uh, either side of the proposition as to whether ChatGPT is overhyped and overestimated or under hyped or underestimated. So let me sort of put my, uh, uh, stake in the ground, uh, which is, I believe it’s overhyped and underestimated. Uh, so that’s my way of hedging my answer. And, uh, of course, ChatGPT is gonna have, uh, you know, lots of implications
Bhaskar Chakravorti: 00:13:07
for, um, you know, for people, for jobs, for institutions, uh, for universities. Uh, like the one we are sitting in right now. It’ll have an impact on many occupations, uh, already, uh, just today. Um, uh, McKinsey, uh, uh, Partha and my, uh, former, uh, employers, old, old employers, uh, put out a big report, uh, which says that,
Bhaskar Chakravorti: 00:13:29
uh, uh, thanks to generative AI and other forms of automation, uh, we’ll have, uh, an additional $4.4 trillion of, uh, global GDP, uh, in, uh, you know, annually, uh, years to come. And, uh, worker productivity might go up by, uh, 3.3%, a stunning 3.3% growth annually. Uh, just given where we are in terms of work worker productivity,
Bhaskar Chakravorti: 00:13:52
that’s, that’s quite a leak. Uh, and McKinsey’s not alone. Other, you know, uh, red Redoubtable analysts such as Goldman Sachs, uh, has a widely circulated report, uh, which you know, pretty much has the similar things. Uh, maybe the numbers are a little bit different. Uh, they talk about 300 million, uh, jobs affected, uh, 7 trillion,
Bhaskar Chakravorti: 00:14:12
uh, impact on global GDP, you know, $7 trillion, $4.4 trillion. We’re talking about trillions. We’re talking about very large numbers, big numbers and big names. Um, so, uh, you know, here’s, uh, here are kind of, uh, I would say five things for us to keep an eye out for. And, uh, so I’ll describe these five things in terms of five gaps,
Bhaskar Chakravorti: 00:14:33
uh, that we need to traverse. Uh, the first gap is what I would call the Yogi Berra gap, uh, allegedly Yogi Berra who said many things on the internet, which, uh, apparently he didn’t say. Uh, but, uh, uh, no one else seems to be claiming this, but he said that, uh, it’s hard to make predictions, especially about the future. So now we think about making
Bhaskar Chakravorti: 00:14:50
predictions about what ChatGPT and, uh, generative AI and large language models are gonna do in the future. Uh, Uh, every time we tried to make predictions in the past about the impact of technology, on productivity and the economy, every single time we have been wrong, every single time we have been wrong. Let me say that again.
Bhaskar Chakravorti: 00:15:10
Every single time we have been wrong. So, if I may make a prediction today, and I’m gonna stand with my prediction, it is that whatever we predict right now, we are gonna be wrong. Uh, so let’s keep that in one second. Is, uh, what I would call the zero sum gap. The zero sum, uh, nature has to do with the inputs that are required, uh, to,
Bhaskar Chakravorti: 00:15:31
uh, uh, to actually grow and build, uh, future generations of generative AI. And those inputs are talent, uh, which are in short supply. Of course, the Gordon Institute is doing a phenomenal job of turning out great talent, but we need many, many thousands of more garden institutes to turn out the kind of talent that’s needed for scalable, uh, uh, applications, uh, or use of AI.
Bhaskar Chakravorti: 00:15:53
Second is data. Uh, of course we are generating tons and tons of data, but the data needs to be accessible. The data needs to be, uh, uh, uh, useful, uh, and and so on, so forth. Third, we need money. Uh, fourth, we need regulations and governance and oversight and ethical reviews and so on.
Bhaskar Chakravorti: 00:16:08
So all these resources are in short supply. So, when generative AI starts sucking up all these resources as they have been in the last few months, it means those resources are taken away from other applications of AI. Such as applications in healthcare, such as applications in crisis response, such as applications in industrial operations.
Bhaskar Chakravorti: 00:16:28
And there’s only a limited amount of the stuff to go around. So my second concern is that our obsession with ChatGPT and generative AI is actually gonna harm AI innovation overall because limited resources are gonna be funneled only in one narrow area as opposed to, uh, broader application than potentially societally more meaningful applications of AI.
Bhaskar Chakravorti: 00:16:51
The third, uh, has to do with the trust gap. Uh, we are already seeing a lots and lots of concerns, open letters and so on and so forth by luminaries, uh, jumping up and down, you know, kind of saying, Hey, we gotta stop this thing. We gotta regulate. We gotta call a six month halt. So a lot of concern about, uh, uh,
Bhaskar Chakravorti: 00:17:07
trust in, uh, what we have unleashed. And not only trust in terms of what will these machines do, uh, when they get out of control, but also trust in the information that’s being delivered as a response to the prompts that we put in already. We know that these, uh, uh, uh, algorithms, they hallucinate, uh, they can give us information
Bhaskar Chakravorti: 00:17:26
that is not entirely reliable. And we know that this is gonna improve over time, but there are many aspects of, uh, the generative AI system, uh, that could potentially get much worse. Uh, we just had, um, you know, uh, Donald Trump once again showing up on the front pages of the newspapers yesterday. Just imagine generative AI, what’s that’s gonna do to the flow of disinformation
Bhaskar Chakravorti: 00:17:48
that’s gonna begin, already began yesterday and go all the way to 2024. I shout out to think of, uh, what our trust gap is, is, uh, how big it’s gonna be. The fourth is the competitiveness gap. Just given the amount of computational power that’s needed for simple prompts, uh, uh, it means that. The entire investment in ChatGPT
Bhaskar Chakravorti: 00:18:06
and generative AI is gonna be limited to very few companies. Very few companies will be able to afford the kinds of computation that is needed, uh, to really sort of, uh, you know, keep this going. And of course, all of that is gonna improve. There’s gonna be a modified version of Moore’s Law that’s gonna apply,
Bhaskar Chakravorti: 00:18:25
but still, this is gonna be the domain of a handful of companies, which means that competitiveness in this space is gonna be challenging. And governments are playing a very limited role in terms of advancing, uh, generative AI and AI in, uh, more, more broadly. They’re pretty much offloading this to the private sector, which is an area of concern.
Bhaskar Chakravorti: 00:18:45
And the five, uh, the fifth area that I’m concerned about is the environmental gap. The amount of energy that’s needed, even for, uh, responding to a simple prompt is enormous. In fact, uh, the computational energy that was needed, uh, for, uh, AI language models, uh, was, uh, in, in January alone, January of this year. Uh, It was equivalent to the
Bhaskar Chakravorti: 00:19:04
energy use of 175,000 people. So, this is an expensive proposition. Of course, that expense is also gonna go down over time. But what that means is we are gonna adding, we are gonna be adding to the climate, uh, challenges that we currently face and taking energy away from other uses. So, with that, uh, all those caveats,
Bhaskar Chakravorti: 00:19:23
I feel we are underestimating, uh, Generative AI and overhyping it. Thank you.
Will Trevor: 00:19:32
Overhyping and Underestimating. Thanks. Uh, thanks Bhaskar. Uh, Partha can we, uh, come to you?
Partha Ghosh: 00:19:39
Yeah. I wanna first thank, uh, Bhaskar and Hanh for laying out a beautiful backdrop. Indeed, uh, as both of you summarized there, not really much to add, except I think it is indeed underestimated. I do not know about the hype, depending upon which quarters you go. Some places overhype, some places under hype, so I would not take a
Partha Ghosh: 00:20:00
position on hyping, but definitely take a position on estimating. Who knows what is a new future that we are going to build. So what I was thinking of taking a step back a little bit and try to look, put it in the context of last million years of development, and I think it’s important that when we are in a points of inflection,
Partha Ghosh: 00:20:21
we understand what is the nature of inflection that we have to deal with. Then are we prepared to deal with the transformational nature of infection? Now, when you take a historian view, firstly you must realize that the evolution of artificial intelligence has happened over time. If you go back to the ancient scriptures, the how the knowledge was structured,
Partha Ghosh: 00:20:46
the strings, the layers, was based on a certain kind of logic, and that is the logic which has edible the evolution of AI as we know today, of course, it is very different. It is transformational in nature. So when you look at AI, it’s a both of a product or transformation. As well as evolution, and that’s where first is different.
Partha Ghosh: 00:21:08
The second thing we have to keep in mind, you know, from the stone age to computer age, to the digital age, to social media age, to the Trump page, as Bhaskar was talking about, you know, the tools that we have developed has been passive in nature. The tools did what we wanted them to do. But with Generative GPT and I think the term GPT is very important, generative pre-train, pre-train transformer,
Partha Ghosh: 00:21:32
it is a very different kind of transformation where the tool itself would be more intelligent than the human. There’s a potential, though I debate that it may not be true, but it is a possibility that this neural network person, when they come together, they will create some kind of intelligent power, which is significantly different from the tools of the past.
Partha Ghosh: 00:21:58
So that’s a very new kind of discontinuity, which we have never been shaped before. So that leads to the next question. If this is the new power, which is coming to play, which we have never dealt before, but we have dealt with discontinuities in the past, what are the lessons we could draw on and what are the newness of this fresh power of ChatGPT?
Partha Ghosh: 00:22:21
That leads to the second point that I like to touch on, which Bhaskar you touched on maybe the, you talked about the five points. I had 10 points, but 10 points, trevor would get upset with me, so I’ll not talk about it. There are lot of different dimensions we have to think differently now. The first point, often people raise.
Partha Ghosh: 00:22:41
What would AI do to society? Any technological advance has generally helped society to move forward. Uh, AI would definitely, Alhanni, were talking about helping productivity, customer service, and all the current business model would fundamentally get more efficient. Or they’ll get transformed. Like I don’t think the healthcare
Partha Ghosh: 00:23:02
in the future would be the kind of healthcare we know today. We will call it human care, where AI would be tracking individuals to keep us healthy as opposed to delivering medicine. But there, that would be the kind of transformation that we be, we should be prepared for in any industry that we are talking about. And the third kind, which I, I
Partha Ghosh: 00:23:24
absolutely agree with both Hanh and Bhaskar, that the future is unknown. What AI, ChatGPT would open up. We don’t know as much as we didn’t know when, uh, James Wat realized the power of steam, lot of different things happened. But this one is no longer the power of steam, which is passing. It is intelligent power. It is significant possibility.
Partha Ghosh: 00:23:53
It reminds me of Tego Tego must will realize. I will not say it in Bengali, but Tego wrote in one place within the finite, you are the in finite. And I think within the finite definitions of the algorithm, AI could open up the in finite possibilities or in another place Tego rose within the limits. You are unlimited shi maji
Partha Ghosh: 00:24:15
to within limited unlimited. So clearly it is difficult to estimate, so I cannot say whether it’s underestimated overestimate, but clearly it would be a new chapter that will open up, which I see very positively. I’m connected with four AI companies, which I sit on the board and I want to share with you what’s going on. One is in the water space, the other
Partha Ghosh: 00:24:40
one is in the agricultural space. The third is in management consulting space, and the fourth one is in blood and analytical space. Similar to what Theranos was trying to do. Now what I see there is the application of AI, which is solving very difficult problems. Water management of the world, evolution of the pr, soil productivity given
Partha Ghosh: 00:25:07
the soil is becoming quite tired and fatigued in different parts of the world. And then we are talking about distributed blood, blood analytics, which is similar to, except this company has a product. In fact, they have got hundred installed base. If any one of you want to invest, this company is ready to take off. And the other one is a company called
Partha Ghosh: 00:25:27
Quantum Logic, which is in management consulting, but they’re not looking at management consulting as Bhaskarcar. And I have known in the world and McKinsey company, you know, we used to spend six months doing diagnostics when I started, which was way before Bhaskarcar. And these days we do it in six days. But in the world of AI, we will do it in, uh, six seconds now.
Partha Ghosh: 00:25:50
So the world is changing. That would be the transformational nature. Now, the final point I want to make, which fortunately both of you have not touched, so you have given me some space, but that’s where I feel the biggest challenge. I think if you look at the development of the human beings that we have basically helped the development of the left brain, which is logical, which
Partha Ghosh: 00:26:19
is knows how to process languages, and that’s what AI would do for us. So the relative value of the left brain would become slightly challenge. What is the right brain, which is imaginative, which is intuitive, which is creative will perhaps become more important, and that’s the only way we should be able to harness the power of AI as well as control the
Partha Ghosh: 00:26:47
power of AI when it comes to ethics, fake news, misinformation, et cetera. So that leads to a very important point that I would like to make for all of us involved in education. That if we continue to educate our students, whether they are in the kindergarten or in the grade 12, or in universities or in higher education, we have to focus not necessarily the
Partha Ghosh: 00:27:11
logical development of their, uh, left brain, but also develop the creative side of the right brain, which means education system, the corporate development of, uh, individuals. The training systems have to be fundamentally different, and I don’t know what it should be looked like, look like, but I could see that there’s a need for more the Albert Einstein re
Partha Ghosh: 00:27:35
tego type of dialogue where one is a scientist, which is a mathematical in nature, but philosopher for iste, go, who, as you know, is a Nobel Prize, one of the first Nobel Prize winners who. Artistic, imaginative and we have to learn how to blend the two in the education system. So we are indeed at a point of infection. We have to think new.
Partha Ghosh: 00:28:00
I will end by one point for particularly youngsters in the webinar, that we would be definitely faced with significant mode, number of choices. So that’s point number one. But these choices, the life cycle of the choices will become shorter cause the world will be changing rapidly under the influence of AI. So to be able to make quick bets,
Partha Ghosh: 00:28:22
which is wider in perspective, would be very important. And then I think I would like to quote and end with a very state, a statement of Nelson Mandela who made, did make, may your choices reflect your hopes, not your fear. You have to be concerned about the fear as Boker was pointing out about Trump. But we have to learn how to deal
Partha Ghosh: 00:28:49
with ’em and put the guardrails, so such possibilities are limited while the possibilities of limitlessness of human development continues. With that, I’d like to thank you very much, you pasa.
Will Trevor: 00:29:05
Um, folks, just a reminder, if you do have questions, I can see some of the questions going into the q and a. Uh, please do push them there. Um, and, uh, if I can now ask James, uh, would you like to, uh, address the topic is, uh, ChatGPT, overhyped and Overrated or Under Hyped and Underestimated?
James Intriligator: 00:29:25
Sure. Thanks very much. Will. Um, and thanks to TGI for organizing this great event and fellow panelists, thanks for coming along. Uh, yeah, so I guess, uh, from my perspective, ChatGPT really is the most powerful tool ever created. Um, if you, if, uh, a thousand years
James Intriligator: 00:29:38
from now, if we’re still around, we look back on human history, I think you’ll see in December, 2022 there was this sudden acceleration in every aspect of human flourishing, whether it’s creativity, productivity, innovation, medicine, um, you know, there’ll be this incredible non-linear, uh, we’re in this period of non-linear acceleration right now where, uh, partly, uh,
James Intriligator: 00:29:59
Because ChatGPT is so powerful and other LLMs and other approaches for machine learning and AI, but I, I would contend in particular ChatGPT, um, it really is helping feedback onto itself. So programmers are making better next gen versions of ChatGPT by using ChatGPT. So everything is kind of suddenly non-linearly accelerating. Uh, if you haven’t been following this
James Intriligator: 00:30:21
kind of on the web and YouTube and the articles, I mean, most of the stuff that. Appears in newspapers, et cetera, is, you know, weeks and weeks out of date. Usually that wouldn’t matter, but the speed at which things are happening now is, is really, uh, astounding. Um, there’s a lot of issues, of course, uh, I, I should say, of course. So in my, my perspective, I think it’s,
James Intriligator: 00:30:40
it’s, uh, chichi BT is underestimated. And in terms of hyped, I’m not sure. I think, I think it’s both over-hyped and under-hyped. Uh, I guess I’d sort of rather it be a little less hyped so that it doesn’t pull in so many, uh, other charlatans, mis users, uh, noise generators, et cetera. But overall, that’s my assessment, uh, in terms of.
James Intriligator: 00:31:03
Uh, sort of philosophical issues. There’s tons of them. There’s been a lot of written about the things about whether ChatGPT has any kind of understanding and intelligence. Is it just parroting back, you know, et cetera, et cetera. Um, I don’t wanna dig into any of those issues, but I’ll, I’ll say from my perspective, I, I kind of take a
James Intriligator: 00:31:16
against shiny and approach and I think that it’s generating symbols that are useful for humans to do stuff with. So whether you wanna call that intelligence or not, it’s definitely helping facilitate further exploration. So, I, I personally would say that yes, it is intelligent. Also, for those who haven’t played around with version four versus 3.5, it’s
James Intriligator: 00:31:36
shocking how much better version four is. So, I. Makes me very excited to see where things will go over the next six months or a year. Um, of course, those are the philosophical issues. Then there’s the ethical issues, which are, you know, huge. Uh, in terms of the, the first question
James Intriligator: 00:31:48
here in our q and a, in fact was about cheating and what about kids in school and all the, the whole model of having kids go home and write essays and turning them in, that just won’t work anymore. There’s no way to catch what is generated by a human or. What is generated by ChatGPT if you’re interested in this, there are tons of, um, recipes that you could find online.
James Intriligator: 00:32:09
You get ChatGPT to write it. Then you go through Grammarly and then you go through. There’s a couple of ways you can kind of play with the text and there’s no way any algorithm can ever tell whether it was generated by a human or a machine. So that’s not so good. Uh, I think it’s gonna really fundamentally change the way we
James Intriligator: 00:32:23
have to educate people in subtopics like English for instance. I think it’s personally, it’s gonna be less about the final deliverable and more about documenting your journey through, uh, the thought process and your journey, uh, by using ChatGPT to explore these interesting, uh, spaces. And that will be kind of more about what maybe English is like.
James Intriligator: 00:32:42
It’ll have to be a whole new, whole new areas of education will be required and frantically it’s in some sense, right? I mean, starting next, uh, semester in the fall, there will be. Tons of essays and student projects that already happen, but that are done purely by ChatGPT. So we need to change the way we think about education, how we think about,
James Intriligator: 00:33:01
uh, training people to use this tool. Again, from my perspective, it’s the most powerful tool ever created. It can do anything. It can impact any field. Uh, it can of course. Also another ethical in I issue is around disinformation, like was raised earlier. Uh, job displacement. There’s lots of things here.
James Intriligator: 00:33:16
Um, longer term impacts, power consumption. Copyright issues is another one that we haven’t explicitly touched on yet today, all of the training was done on a great deal of material that was, uh, copyrighted, for instance, and know is it right that it was used? Is it right that it was ingested or digested?
James Intriligator: 00:33:33
From my perspective, personally, I think that it was a crime. Uh, and it would be nice to find some way to kind of. Create, I dunno if it’s reparations, that’s probably too loaded of a term in this context, but there needs to be some way to make that right. But at the same time, I don’t think we should throw out the tool that
James Intriligator: 00:33:48
was built by this, uh, well, I’ll say crime, uh, cuz the tool is so powerful as force for good as well. Uh, other issues are around sort of bare work and pay income inequality, right? So the uh, if this is gonna become another tech tool that serves to increase, further increased income inequality. It could very well do that. People who know how to use it and can
James Intriligator: 00:34:09
use it effectively within business domains will become much wealthier and others will be lacked aside. So that’s another issue that we need to think about. And of course, there’s huge impacts and issues from a social justice perspective. Uh, but so those are sort of the concerns in, from my perspective, I think that it’s a fabulous tool that can really help
James Intriligator: 00:34:26
anyone do anything they want to do better. So if you’re a surgeon, it can make you a better surgeon. If you’re a musician, it’ll make you a better musician. If you’re an engineer, you’ll be a better engineer. Anything you want to do, it can help you do better within reason. Right?
James Intriligator: 00:34:40
I mean, if you wanna be, uh, bodybuilder, it can’t, well actually it can’t, it can give you wonderful, uh, workout plans and health plans, eating plans. Um, uh, In terms of, I’ll just say a few words on kind of how I think of ChatGPT. You know, you, you often hear people complaining about how it hallucinates and it just makes things up, et cetera. Uh, from my perspective, that’s
James Intriligator: 00:34:58
part of the beauty of it in a sense. Uh, and as long as you understand that you can actually leverage that and use it to your own advantage. So from a, so as a professor sort of of human factors engineering, part of what my field looks at is how people, uh, what mental model they have as they interact with a piece of technology. And I think that open AI has done sort
James Intriligator: 00:35:16
of a disservice to the world by putting a search, like front end interface on it. As, as users, we tend to see a search box and think, oh, search engine, and we just go about doing search. And it’s really nothing like a search engine. Uh, it’s, it’s unfortunate that they did that because it really misled many people, uh, who then.
James Intriligator: 00:35:36
Judged it inaccurately as, oh, it’s not a very good search engine. Well, it never was. It, the, the way I think of it is it’s really sort of a, um, I have my own little mental model, but it’s a, a, uh, text generating glider, a transversal glider that lets you fly through trillion dimensional spaces and explore those spaces.
James Intriligator: 00:35:54
Sort of a what if engine, how might we possibly explore kind of engine? And the key to using it well is really, uh, understanding that, that it’s not a search engine, it’s a glider and putting it at the right point in space, aiming it in the right direction, and having to glide. Uh, and so you can do all kinds of stuff if you think of it that way.
James Intriligator: 00:36:12
Um, uh, again, just a couple of. Uh, points and examples. So one thing, if you have used it, uh, you probably already have bumped into the idea that it’s much better to ask for multiple answers. So give me seven ways I might design a new product for this toy company. I designed a bunch of products in the last couple months using ChatGPT.
James Intriligator: 00:36:32
Um, can you identify ways that this products could be sold to our customers? You know, tell me about our customers chat sheet. BT knows all about, let’s say, the particular customers of a company. It’ll tell you about the customers. You can say, give me seven clusters of customers. Name each cluster.
James Intriligator: 00:36:47
Give me a description of each customer group. It’ll do that. You can then say, okay, I wanna sell my products to some of those customer groups. I wanna hold an event somewhere in Boston. Tell me where I should hold the event, what the event should be like, what day of the week, and what time of day the event should happen
James Intriligator: 00:37:02
to maximally target those people. Now it’ll do it. It’ll do it. It might be making it up, but if you look at it and if you consider, you know, you never would just copy and paste. You sort of take what it says and you, you consider it anyway. So, uh, That, that’s just one example. I, I wanna just mention a few of the
James Intriligator: 00:37:15
other ways that I’ve used it, um, myself over the last month or so. So I’ve been doing things like designing behavior change, interventions to reduce cancer rates. It’s fabulous at doing that. Uh, I’ve been designing a new course for kids to teach them cybersecurity. Done fabulous work at that. First outlining what kids nowadays
James Intriligator: 00:37:32
know about cybersecurity, what they don’t know, where the gaps are. Um, I’ve been doing new methods for literary criticism, uh, doing some fun work around customer segmentation. Back to the thing about tell me about eight kinds of students I might have in my class. It’ll do it. And if I can give a four question
James Intriligator: 00:37:48
questionnaire to my students to classify them into different segments, what would those four questions be? It’ll do it. Um, I had it the other night, write a better ending for the final season of Game of Thrones or right this final season of Game of Thrones. Much, much better than the one that actually came out.
James Intriligator: 00:38:05
Wouldn’t, it’s not very difficult to get much better than that one anyway, but that, anyway, uh, engineering, it’s been using, I’ve been using it to design machine learning algorithms, to write code, to do the machine learning things, to take, um, data and process it in novel ways. Harvesting energy from oceans, been using it for lots of kind of, Um, green
James Intriligator: 00:38:23
economy issues and also for DEIJ issues. So I’ve been, uh, giving it my courses and saying, you know, how might I incorporate DEIJ themes into this class? From an educator perspective, it’s been fabulous in helping me identify opportunities for bringing DEIJ issues, uh, diversity, equity, inclusion, justice issues into classes, uh, songs, poetry, you know, it’s been doing
James Intriligator: 00:38:42
kind of everything you can imagine. Again, from my perspective, it can really help anyone do anything better as long as they understand it’s not a search engine and they can kind of fly it in the right ways.
Will Trevor: 00:38:52
Alright. Thank you, James. Lori, if I can call upon you, um, is this ChatGPT overhyped and overrated, or underhyped and underestimated?
Lori Mazor: 00:39:03
Absolutely. Um, again, thank you for inviting me and I come here to join this esteemed panel as the street artist among you, I have for 36 weeks been using generative AI, starting with, um, mid journey I, uh, application for image generation. And so by the time ChatGPT came out, I had really been experimenting in a very vernacular way about what this technology
Lori Mazor: 00:39:27
means for how we engage, uh, humans and machines, uh, without a computer science background, but really thinking about it from the perspective of how do we balance. Chaos and control. And I think this is potentially the most brilliant aspect of generative systems as opposed to discriminative systems, which are the kinds of technologies that we’ve been accustomed to and that we
Lori Mazor: 00:39:56
actually feel quite comfortable with. And that, um, allow us to kind of proceed with digital tailorism, which is kind of the underlying mechanism of automation and moving things from human processes to machine processes. And I think we’ve seen what some of the downfalls are there. But what I wonder and why I think the technology is underrated is if we
Lori Mazor: 00:40:23
could actually embrace chaos as humans. I think that we are going to see that this is going to be one of the greatest challenges of our time. It both excites us and scares us to the very core. I wanna use that analogy that you talked about when you said hallucination and this idea that we need to be able to harness something or control it or
Lori Mazor: 00:40:50
bring it into a set of social norms. I think that this is actually one of the fundamental problems with our misunderstanding of creativity and the way that we train children and ourselves In many ways, I think children are born with infinite amounts of creativity. And I’ll use an example in my own life of going over to a neighbor’s house at the age of five, and I told a story, I don’t
Lori Mazor: 00:41:13
even know what the story was, but I came home and the neighbor told my parents about the story and accused me of lying. Now at five, I didn’t know that what I was telling wasn’t true. My boundary between imagination and reality didn’t quite exist yet. And I think this is where we found the very first samples of generative AI. And as I’ve watched it evolve over this
Lori Mazor: 00:41:42
very short period of time, I’ve seen us try to put it in a box to try and get it to perform for us to behave the way that we ask children to behave. And so I think we’re in a very important moment where it hasn’t quite jumped the shark yet, and we have an opportunity to have a really critical conversation about how we embrace chaos. And there’s a really
Lori Mazor: 00:42:08
interesting factor that’s called temperature in any AI system. It’s chaos when you’re talking about, um, image generation and it’s temperature when you’re talking about text generation. And it has to do with how much. Randomness we allow in the prediction of the next word. So for example, a low temperature setting, if you say I love, will always generate
Lori Mazor: 00:42:35
the word you because you is the most popular word that’s used after I love. But if you allow that temperature setting to be relaxed, if you dial it up to one, it might say, I love chocolate, or I love animals, or I love my mother, or I love books. Right? And that is where I think we have the potential for innovation.
Lori Mazor: 00:42:56
And I think that’s actually one of the most important parts about how we can use this for business. So yes, there are certainly ways that we can improve productivity, but I would challenge us to think about how we actually invite chaos and how we use that in a way to solve some of our biggest problems.
Will Trevor: 00:43:17
Thanks, Lori. And, uh, thanks to all the panelists for that, uh, for that first round of some very full and, uh, interesting answers. Um, thanks also to those of you in the audience who’ve been posting your your q and a, uh, questions. Uh, I see that, uh, James is also typed an answer into to some of the questions and panelists, if you see a question
Will Trevor: 00:43:36
in there that’s directed to you, do feel free to to, to type in an answer. Um, the first question, uh, I’m, I’m going to pose. Uh, James has touched upon this, um, uh, BAS path. Uh, Ville touched upon it as, as well in, in terms of, uh, education. Um, in terms of talk about hallucination, I was at a, a webinar a few weeks
Will Trevor: 00:43:59
ago with the librarians from the Tish Library, uh, and they, they said that they, they have students coming in, uh, now, um, with, uh, references that have been suggested by, uh, ChatGPT, that are themselves entirely, entirely fictitious. Um, there’s, there’s questions on education that come in from, uh, Laura Lee, and this will address your question, hopefully, uh, and Stacy as well.
Will Trevor: 00:44:22
Uh, it’s also a question, uh, that since the pool is there, uh, I decided to, to generate with the help of, uh, of ChatGPT. So the, the this, uh, first ChatGPT, what would we ask you? And, and here’s, here’s what, here’s what it gave us. So, uh, in what ways can ChatGPT transform the educational landscape and other potential benefits and challenges?
Will Trevor: 00:44:44
How might ChatGPT impact the role of teachers and students, uh, for example, and are there any concerns regarding the overreliance on ChatGPT in the educational settings? Um, Bhaskar, do you wanna tackle that one first?
Bhaskar Chakravorti: 00:45:01
Yeah, sure. I, I think, um, uh, you know, ChatGPT, um, is a tool, uh, like so many tools that we’ve had in the past. And, uh, it is an advanced tool as you know, uh, many of my colleagues here, uh, have pointed out. It allows us to, uh, you know, extend our ability to, uh, write our ex our, our ability to, uh, create visuals.
Bhaskar Chakravorti: 00:45:24
Our ability to, uh, imagine, uh, extends it way beyond, uh, you know, what our current capabilities are. So in terms of our impact on education, uh, it definitely, uh, will enhance our. Ability to educate, provided we, uh, on the faculties of universities like Tufts and all the other universities, and not just the universities, the school systems and so on, are prepared,
Bhaskar Chakravorti: 00:45:46
uh, to embrace this tool and, uh, and then figure out how do we now enable our students to become better learners and to be able to produce, uh, uh, better outcomes, uh, of that, uh, of that education in combination of the tool. So I know that there’s some questions there which, uh, have to do with, uh, might this stifle creativity, might this lead to greater cheating?
Bhaskar Chakravorti: 00:46:12
Uh, might this, uh, you know, give students, uh, uh, uh, access to false information? The answer is yes to all of that. And it’s up to us as educators to anticipate, uh, those traps and then, uh, guide our students and ourselves, uh, to skip over them. So I think, uh, the bar has been raised and, uh, it’s up to us to
Bhaskar Chakravorti: 00:46:33
try and, uh, and, and, and meet it. Uh, it’s not gonna be easy, uh, because the technology keeps moving faster than our ability to comprehend it. Uh, but I do believe that over the long haul, uh, it is gonna be, uh, uh, an accessory to education. It’s gonna be an accessory to knowledge. Um, I do wanna make one comment on the importance of ChatGPT and generative AI
Bhaskar Chakravorti: 00:46:57
relative to AI Overall, I feel that in many of our, uh, observations, uh, we have been combining those observations. Uh, and I do feel it’s important to make some distinction between generative AI and other applications of AI. And then when we talk about what are some of the biggest problems that we want to solve for. Uh, of course, uh, we would like to
Bhaskar Chakravorti: 00:47:17
solve for problems of, you know, getting knowledge, uh, you know, giving, uh, uh, uh, people, uh, the ability to, uh, uh, you know, uh, uh, fix things or, or, or write an essay and so on. Uh, but we already have several tools, uh, that allow us, uh, uh, to do that or allow us to educate children to do that. If I think about problems to be solved, um, one problem that rises
Bhaskar Chakravorti: 00:47:43
to the very top, uh, in terms of societal needs, uh, to me is the application of AI and healthcare. And, uh, we’ve just come out of three years of, of pandemic, uh, where, uh, we did not even have the tool to tell whether somebody had covid or not. You put in a test and it took three weeks at the heart of the, uh, of the pandemic, three weeks for a result to come back.
Bhaskar Chakravorti: 00:48:05
There are multiple AI teams that tried to come up with an algorithm that could catch covid from the sandor cough or something like that. Almost all of them, not almost all of them failed. So I really feel one of the biggest problems to solve for us as a society is to save lives and to improve lives. And my worry is generative AI may be
Bhaskar Chakravorti: 00:48:26
sucking up a lot more resources away from those kinds of applications. So that’s also an aspect of education that we need to think about.
Will Trevor: 00:48:37
Okay. Thank you. Thank you, Bhaskar. Um, Hanh, same question to you. How, how might we be able to use, um, Generative AI as a, as a tool? How might it transform the educational landscape? What are, what are perhaps some of the, uh, the challenges that you see?
Partha Ghosh: 00:48:53
Well, could I add point to the respond to Bhaskar’s very thoughtful points that I feel the whole relationship between faculty and students have to change because the world of AI and the world of generative AI. Would provide us with the information. The information exchange will happen because of this tool, but the, both the faculty and the student has to come to
Partha Ghosh: 00:49:21
the learning space, both as students. The role of the faculty would become a guiding, the process of managing the knowledge transfer as opposed to coming as I, I know it all, you have to learn from me. Because that would definitely compromise the, all the value that, uh, generative AI could offer. Whereas the role of the faculty in the
Partha Ghosh: 00:49:48
future would be to unlock what I think. Uh, Lori, you were talking about the ability to deal with chaos ness, imagination, and that is a very different kind of, I would say, teaching process. If we use the word teaching, maybe the word is not teaching. The word is unlocking that inner creativity. So that’s a very different kind of guiding
Partha Ghosh: 00:50:15
system, if I may use the word, as opposed to professorial system where the professor comes with all the knowledge that he knows and he imparts it on student. That era is over. So, if we really want to do justice to the power of AI, justice to humanity, uh, we have to engage these two powers with a very different mindset.
Will Trevor: 00:50:44
Thank you, Partha. Hanh, if I could bring you in at this point, um, what, what do you, what do you see as perhaps the, the uses of this tool and perhaps the challenges as well?
Hanh Brown: 00:50:53
Sure. I’d like to echo all that’s been said in terms of education and the impact of Jet c p t. One more thing in terms of. Let’s say elementary, middle school, high school and college. I think it’s crucial to have an actual curriculum, a class chat, T p T for elementary, whatever that entail, but
Hanh Brown: 00:51:12
present it in such a way that kids at that age use it responsibly and not perceive it to be something frowned upon. Because here’s the thing, as much as I like to say social media is bad for kids, and I did when my kids were younger, they’re 26, 23, and 21, but that was my exact sentiment. Social media was bad for elementary, middle, and high school kids.
Hanh Brown: 00:51:39
Well, that didn’t go well, you know, during those years. So I see chichi PT is gonna be integral to our home life, you know, workplace. So I think it’s best to package it in such a way that elementary kids can use it wisely. And teachers perhaps even encourage them to learn and grow and not look at it as something frowned upon.
Hanh Brown: 00:52:03
So that’s one point that I wanted to add. Now, as far as education for older adults, well, I come from the aging sector, so I always wanna add a spin, how that would impact the aging population. Well, this transformed education as we mentioned, by providing personalized learning experiences tailored to a student or an older adult pace and learning style, right?
Hanh Brown: 00:52:28
It also enables 24 7 learning offer immediate responses, and this is very, in my mind, relevant for lifelong learners among the aging population as they are encouraged to continue their education. As we age. So that is one of the key components for thriving in life. Now, I understand there’s issues over-reliance on AI could lead to
Hanh Brown: 00:52:58
decrease in HU human interaction and perhaps even risk of widening the digital divide among those who are not tech savvy for some older adults. And I think we should shift towards facilitating learning and using AI as a tool, but not necessarily the primary mode for instruction. Now for older adults who may have access to human facilitators that can help
Hanh Brown: 00:53:28
them make, who would need access to human facilitators to help them make the most out of AI educational tools. So I think that’s really important. The proper training. View it as something integral to your life and innovate. And not something to be frowned upon. Now, recognizing there’s guardrails, of course, with everything mentioned here.
Hanh Brown: 00:53:51
So that’s the two points I wanted to add.
Will Trevor: 00:53:54
Okay, thanks, Hanh. Um, Lori, if I can come to you, um, again, uh, you, your, your background is in, in education. Uh, how do you see ChatGPT being, being utilized within the educational landscape?
Lori Mazor: 00:54:07
Yeah, the, the aspect of it that I found find most interesting is, um, the possibility for inquiry-based learning, right? So being able to self-direct your educational path, um, and having access to a breadth of knowledge. Um, and I think it really plays to the polymath. So I think that we have up until this
Lori Mazor: 00:54:25
point, really, um, as much as we talk about interdisciplinarity in education, we still are very disciplinary. And, and the truth of the matter is as you move from education into the work environment, to, we are extraordinarily focused and specialized. And so all the talk about range in education actually doesn’t get applied. And so what I really love thinking about
Lori Mazor: 00:54:49
is what if every person were their own? CEO? What if every person could kind of span the breath from science to math, to art? Um, and instead of having to be experts in one area, if we can draw upon this compendium of human knowledge, um, it doesn’t mean that there will not be a place for experts, but I think there’s more of a place for people who might
Lori Mazor: 00:55:14
be able to cross and span those gaps, which I think is really what starts to help in solving those problems, right? Connecting things that have been, uh, disconnected.
Will Trevor: 00:55:30
Thank you. Uh, thank you, Lori. Um, I’m gonna take a, a slightly different direction now. James. I’ll come to you and then, then I’ll come to you. Um, Couple of questions have appeared in the, in the q and a so far, uh, talking about the, the potential for, uh,
Will Trevor: 00:55:44
disruption in terms of the, uh, employment market and, uh, and, and job losses. Um, so the question I have is, is how might chap GPT affect the, the, the job market and employment, uh, opportunities in different sectors? Uh, will it, will it have, uh, for example, a significant impact upon productivity and efficiency in industries? And if so, how?
Will Trevor: 00:56:10
But if we could, if we could focus on that, that sort of job loss, that change in employment, uh, James?
James Intriligator: 00:56:19
Sure. It’s a nice, easy question. Uh, yeah, that, that’s a complicated one. I, I think there’ll be huge changes in the, the landscape of jobs over the coming couple of years, you know, has, has happened previously with sort of the information economy, et cetera. Whole new categories of jobs will open up already there.
James Intriligator: 00:56:36
There are now, um, incredible numbers of opportunities for people who are professional prompt engineers, right? That’s now become a whole job category. Uh, and I think there’ll be many more things like that. People who can effectively use ChatGPT. Uh, so there’ll be new, new types of jobs appearing. Of course, uh, many of the
James Intriligator: 00:56:52
older jobs will disappear. So, you know, it, it’s already been happening for the last, well, forever, basically. As, as, um, automation has been coming around, right? There’s no longer, uh, almost any need for loan officers, for paralegals, for entire categories of, uh, professions aren’t disappearing,
James Intriligator: 00:57:10
have been because of automation and machine learning and, uh, language. Uh, Easy template, matching, things like that. This is adding a whole other layer on top of that. Uh, there’ll be many more jobs that are both threatened but also accelerated by this, right? So if, uh, if you’re a screenwriter,
James Intriligator: 00:57:23
yes, you should be a little bit afraid. You might be replaced by a ChatGPT model. But more likely what you should be doing is figuring out how to use ChatGPT to become a better screenwriter. Uh, if you’re a poet, that’s awesome. The Chachi bt most of it’s poetry isn’t great, but if you play with it the right ways and you kind of move it in the right spaces, it
James Intriligator: 00:57:43
can write some phenomenal poetry. Um, it hasn’t quite brought me to tears, but almost, I mean, it is, it can do incredible things as, oh yeah, if you’re a poet, yes, maybe reject it. Be a Luddite, write your own poetry, rely on your own creativity, that’s awesome. Or embrace it and figure out ways to use it to, to your advantage. So I think it’s gonna be a real,
James Intriligator: 00:57:59
uh, period of transformative growth, contraction and mutation. I don’t have much more to say than that.
Will Trevor: 00:58:10
Alright, thanks James. A, a period of growth contraction, um, uh, Partha, over, over to you. Um, how, how do you think, uh, ChatGPT might, uh, affect the job market and employment opportunities in different sectors?
Partha Ghosh: 00:58:23
You know, it’ll, uh, obviously affect dramatically, but again, I would like to draw on history. If you look through the different revolutions that I’ve gone through, the industrial Revolution revolution, then we talked about the knowledge revolution. Now we are into intelligence revolution. So intelligence revolution is different in the client for the, the first time
Partha Ghosh: 00:58:43
you’ll see the knowledge workers, the people like you and me would be challenged cuz much of the work would be done by ChatGPT and different components of AI. So obviously there would be a significant dislocation as to how the knowledge workers would be engaging with society. But that said, I do feel, as we have seen with the first stage industrial revolution, the steam
Partha Ghosh: 00:59:08
engine, the second stage with the electricity, then the laptops, a empty number of new jobs will be created. There will be different in kind. I see five different industries being born. Each would be trillion dollars plus like one where I will call the green earth industry, which would be all about carbon capture, all about renewables and
Partha Ghosh: 00:59:33
AI would be the fundamental of the core to sense opportunities and the sense problems that the space travel, which will be the second industry, which will double up dramatically powered by AI. The third would be the spiritual machine, which has not been served as yet. We have served our intellectual requirements, physical requirement in the next 50 years.
Partha Ghosh: 00:59:57
It was see spiritual engines. So there have been lot of different jobs. So, We created. But what I worry about what, Lori, you touched on that how would we shape the development of human beings? Because all of a sudden, intelligence is all around, which is, we call it artificial, but becoming more and more natural.
Partha Ghosh: 01:00:20
So how to integrate natural intelligence with artificial intelligence and AI also, also at a very early stage of development, the power of AI in 10 years time would be not by few percentages, by different factors would be significantly stronger. So I think the, the key challenge would be education and how do we educate on one side professional knowledge workers to become more relevant?
Partha Ghosh: 01:00:49
Because as they get displaced with their current jobs, what would be the new jobs that there be hundreds of them? How do we train them fast? And the other side, how do we prepare the young. Children, the people in currently in universities and colleges to manage the interface between ni the natural intelligence and artificial intelligence.
Partha Ghosh: 01:01:12
So in terms of opportunities, there will be millions of opportunities, but it would require different mindset, different, uh, reference points. And as I talked about, even in the education system, the faculty would not be the kind of faculty we have seen in Oxford and Cambridge in the past. There would be faculty which are humble servants serving to unlock
Partha Ghosh: 01:01:35
the imagination of students. And that’s a different kind of a ballgame, which we have to learn.
Will Trevor: 01:01:41
Thank you, Partha. Uh, Lori, I’ll come to you on this, uh, topic and then perhaps we can switch topics in a moment, Baskar, and I’ll come to you about, uh, democracy and, and participation implications of generative AI. But, uh, uh, Lori, what, what are your thoughts in terms of change, changes in employment and skills
Will Trevor: 01:01:56
and potentials for job losses?
Lori Mazor: 01:02:01
Well, I think this is gonna be a bit of a tug of war, to be honest. And actually the screenwriting example is a perfect one, and I’m watching that case study in progress. I mean, we have a union that is, um, Really, you know, at a point of negotiation. But the core issue is really how much we are going to allow artificial intelligence
Lori Mazor: 01:02:20
to play a role in the process. And we haven’t had enough time, frankly, for people to experiment with this technology and understand it, to learn it, to absorb it before we’re. Starting to take a stand. So I think in the United States, um, we are going to say, see a real tug of war with unionization. And, and I think as we’ve
Lori Mazor: 01:02:39
even seen in academia, u unionization has been on the rise. Um, and so there is a lot of protection, um, in, in that I think the first jobs to go, we are going to see overseas. And I, so I think there’s going to be a global impact, um, that frankly we might see as a benefit initially to American businesses. Um, but that will have a, um,
Lori Mazor: 01:03:00
a pretty dramatic effect on how the globe, um, engages. And, you know, just the simple, um, choice to use a chat bot instead of to use people who might be in India or in Pakistan. Um, is, uh, it. It’s profound. Um, I, I think the other labor issue, which we, we can’t overlook is, you know, this technology is profoundly
Lori Mazor: 01:03:25
troubled and very problematic in terms of the way that it’s come into being. And the, um, amount of labor that’s gone in, uh, that was not unionized, that was unprotected, the working conditions of those people, those are things that are well-documented and frankly, those are the harms that we often brush under the rug. Um, but I think we, we do need to take into serious account.
Lori Mazor: 01:03:51
So from my perspective, this is really about thinking about this from a global perspective, um, rather than through the lens of what’s gonna benefit American companies and, and frankly the, um, the stock market.
Will Trevor: 01:04:04
Thanks, Lori. Uh, Bhaskar.
Hanh Brown: 01:04:05
Let me, if I allow, um, be allowed to add
Will Trevor: 01:04:07
Sure.
Hanh Brown: 01:04:08
a spin to this from, let’s say a Baby Boomer point of view. Well, I think obviously ChatGPT, very powerful. It enhances human capabilities, automate mundane tasks, freeing up for older adults who do more strategic roles. Like we mentioned, customer service you can tackle on basic increase, leaving more complex cases for human intervention.
Hanh Brown: 01:04:33
So, this could make jobs more appealing for older adults, allowing them to leverage their wisdom and experience in oversight, problem solving roles. So I think that it will create opportunities as long as folks, older adults adapt and also have a means to learn and be trained properly.
Will Trevor: 01:05:01
Thanks, Hanh. Um, Bhaskar, if I can just change, um, direction, uh, slightly, um, in, in terms of regulation, in terms of, um, politics participation. What, what, what do we, what do you see as the potential risks or benefits that ChatGPT poses to democratic systems and, and political participation? Uh, are there any regulatory
Will Trevor: 01:05:25
or, or governance measures? We’ve talked about guardrails, um, uh, a couple of times now, uh, needed to mitigate the potential prob political misin, uh, manipulation or, or indeed, you mentioned, misinformation, uh, earlier around ChatGPT?
Bhaskar Chakravorti: 01:05:43
Yeah, of course. Uh, really important question. Uh, will, uh, let me, uh, let me touch on two things and, um, one is the need for regulation and oversight. Clearly there is need for regulation and oversight. The problem is, Who’s gonna do the regulation? Because if you have a Sam Altman, um, CEO
Bhaskar Chakravorti: 01:05:59
of, uh, OpenAI, uh, going to Congress and asking, uh, the members of Congress, uh, or other policymakers or regulators to regulate him and the industry he knows fully well, Sam is a very smart man. He knows fully well that none of the people sitting across the table have the ability to regulate a technology that they fundamentally don’t understand. So I feel a lot of the AI gurus and AI,
Bhaskar Chakravorti: 01:06:25
uh, leaders who are asking regulators to regulate them know fully well that they won’t be a follow through. They’re basically covering themselves, uh, with plausible deniability, uh, kind of saying, Hey, we told you that we are creating a frank Science monster, uh, regulate us now. And if the monster is unleashed, then it’s on you.
Bhaskar Chakravorti: 01:06:50
It’s not my fault. So that’s one. I think, and we are living through this moment right now, right? Because you think about members of Congress, forget about members of Congress who are, uh, you know, uh, uh, uh, uh, they’re, they’re, they’re much further removed from things like, uh, AI and so on than, uh,
Bhaskar Chakravorti: 01:07:06
the members who are attending this particular, uh, this particular debate. Even we don’t understand. I mean, I, I, I, I won’t be able to open up, uh, the, the hood and try to figure this out, or I’m, I can’t speak for my colleagues. This, this is a technology that is very, very sophisticated, running very far ahead of our ability to, uh, create guard rail.
Bhaskar Chakravorti: 01:07:24
So that’s an ongoing issue that we need to recognize the fact that, uh, you know, we need to take this much more seriously than we currently are. And just signing a letter, an open letter saying, you know, a six month hiatus, give me a break. That is just nonsense. That’s just being naive. I mean, none of these people
Bhaskar Chakravorti: 01:07:39
are gonna take a break. Google’s business model is being threatened right now. You think Google’s gonna take a break? Of course not. Microsoft is gonna take a break. Of course not. So the, the, and these people live in, uh, Riyadh. That’s one.
Bhaskar Chakravorti: 01:07:51
The second is what I really worry about, worry about is the state of democracy. And the state of democracy, of course, is being unleashed on a daily basis, or the attack on democracy being unleashed on a daily basis. And because of the abilities to be creative, ChatGPT can allow us to be much more creative in terms of disinformation. And that’s only gonna click,
Bhaskar Chakravorti: 01:08:08
uh, going up between now and, uh, 20 and 20, 24 rolls around. Now, here is the issue that I think we are not paying sufficient attention to, which is that democracy comes at a cost, and part of the cost is preserving democracy in the United States comes at a cost of democracy in the rest of the world. Here is why 2024 is a year when we are gonna have major elections
Bhaskar Chakravorti: 01:08:35
here in the United States. 2024 is also a year in which India, Indonesia, South Korea, a dozen African countries, at least seven Latin American countries also have elections. Now put yourself in the place of Facebook, which has just fired tens of thousands of people. Many of them who are in positions of doing content moderation.
Bhaskar Chakravorti: 01:08:57
They’ve also fired their content moderation. Outsourcing teams, limited resources for content moderation at Facebook is all gonna be focused on the United States, right? Uh, historically, 90% of Facebook’s content moderation was focused on the us whereas 90% of Facebook’s audience was outside the us.
Bhaskar Chakravorti: 01:09:17
Now that disproportionate allocation is only have, uh, got, got worse, and the ability of disinformation creation has just magnified because of chat shaping. So if you are worried about democracy, Yeah. Uh, sadly we ain’t seeing nothing yet. And the problem is when we say we, it’s not we in the United States, it’s we in India, we in Indonesia,
Bhaskar Chakravorti: 01:09:37
we in south a, uh, South Korea. We in a dozen sub sub-Saharan African countries where there’s nobody who’s gonna be watching the content that’s gonna be running rampant in those places. And I really, really worry about the state of democracy, uh, both here and in the rest of the world.
Will Trevor: 01:09:54
Thanks, Bhaskar. James, can I come to you on, uh, on this one? Um, and then I’ll come to you Partha, uh, and then I think, uh, we’ll, we’ll take a slightly different tank after that. But, uh, James, what, what, what do you think the implications are for, um, democracy and for, for political participation?
James Intriligator: 01:10:11
Uh, I, I don’t really think I have anything to add beyond what my colleagues have been saying. I mean, I think there are, I guess, you know, there are huge dangers, presence, the, the kind of ability to generate, uh, misinformation, false text. Uh, the other thing that we haven’t touched on too much directly is just the ability to write text that really
James Intriligator: 01:10:28
speaks directly to a target audience. So if you really want to have a message that gets to a very specific group of people, I either for good, right? So, you know, hey, you should really show up to the polls because otherwise, The evil guys will win whatever it is, that’s great. But it could also be, you know, text that is generated to specifically speak to
James Intriligator: 01:10:47
people to make them, uh, believe that the election isn’t, has been delayed by a day. There’s all kinds of ways and, and ChatGPT is fabulous at coming up with texts that really speaks to and motivates, uh, uh, specific target demographic. So, not only will it be creating false information, it’ll be creating false information that’s even more effective, believable,
James Intriligator: 01:11:06
and targeted for a specific group. So I think that that’s the other kind of side of it. It’s not just gonna add noise, it’s gonna add very targeted, strategically defined kind of viruses that it can put out, uh, that people can use to put out and influence elections and really undermined democratic processes in, in many different ways.
James Intriligator: 01:11:23
So yeah, that is terrifying.
Will Trevor: 01:11:25
Thanks, James. Partha?
Partha Ghosh: 01:11:27
You know, I will take a little bit of a different view, uh, just to, um, enable a debate because after all it’s a debate. Uh, on while one side. I agree with both of you that. Indeed democracy is being challenged and rightly so, I would say, because many ways you could say as democracy delivered, you know, Baskar, you talked about India.
Partha Ghosh: 01:11:51
India has been democratic country for last 75 years. But when I look at the performance of India, whether they think in terms of pollution, corruption, nutrition, GDP per capita, and compare that with other countries in the neighboring areas, which are not democratic, and India’s not done well. It’s true, similar, if you look at the
Partha Ghosh: 01:12:13
United States, it is true that the gap between the rich and poor has increased. There are a lot of, a lot of angry people for whatever be the reason. So here is a new force that we have been exposed to, which is the power of AI. So the question to ask would be how the ChatGPT bt, or the power of AI. Would be used to create a governance process of nations governance process
Partha Ghosh: 01:12:44
of the world, which is keeping up with the technology as opposed to worried about the technology, which means maybe there are another kind of crazy in autocracy we lived through before that there were the kings and emperors democracies, relatively new. We have experimented with democracy as Churchill is to say “In absence of any other process, this is the worst of all.
Partha Ghosh: 01:13:06
Could be the best of all.” So now, the question is, could we be also be innovative to use the power of intelligence, both natural and artificial, to create new governance processes, new ways of people who could participate, which could be Democracy plus one, or Democracy 2.0, or some other name, which we don’t know. So I would not get locked up that
Partha Ghosh: 01:13:28
because of AI, because of ChatGPT. The democracy’s challenge. Of course it is challenge, but then we have to be innovative enough to embrace this power to create the next S-curves of ma management of nations, management of society, management of the global governance. Because you, you could also see this is my favorite subject, by the way.
Partha Ghosh: 01:13:53
I can go on for three hours, but I would like to make another point. You know, we have created United Nations after the second World War. We created the League of Nations after the first World War. Both of them are not delivered. We have more wars today. Look at what’s going on in the middle of Europe.
Partha Ghosh: 01:14:09
Let’s look at another Korean peninsula. Look at what’s going on in the South China Sea. What’s growing a crisis and multiple problems in Africa and Latin America. So what has United Nations dev delivered in the name of democracy? Very little. So I think we have time has come to use and embrace this power.
Partha Ghosh: 01:14:28
To create a new humanity, new civilization that the world could celebrate, where every child, the 8 billion people could participate in development of a civilization where everyone enjoys the minimum dignity gracefulness than the world deserves. And that would need a different kind of governance, not the democracy, which has partly worked, but largely did not work.
Partha Ghosh: 01:14:53
And we have to be honest about it.
Will Trevor: 01:14:56
Thank you, Partha. We’re gonna take a, a slightly different, uh, tank, uh, now, and I’ll come to you Hanh, uh, first, and then then to you Lori, if I may. Um, We talked about some of the disruption, we talked about the potential for, for job losses. So, uh, those are, uh, those in the audience, uh, now thinking, well,
Will Trevor: 01:15:15
how is, how is this gonna impact me? And, and how can I, how can I prepare? Um, what, what do you think, uh, in, in terms of individuals and organizations, how can they prepare for this, uh, disruption? Uh, are there any specific skills, uh, or areas of expertise that are become increasingly more valuable, uh, in a world where ChatGPT
Will Trevor: 01:15:37
is more widely, uh, utilized? Um, how can, how can people, how can people prepare, re-skill, retrain, Hanh?
Hanh Brown: 01:15:46
Well, the folks, let’s say, who’s calling in, I’m guessing your college age. My advice is keep learning, creating and innovating tech courses. I know Google has offers some free courses. There are courses on Corsica that’s run by, um, ING, I believe, uh, that’s the last name.
Hanh Brown: 01:16:08
He came outta Google. There’s abundance of some free education ranging from some basic understanding to very deep understanding of machine learning, python and have an open mind. And of course, in learning, proceed with caution. And I would say because let’s say your college age, you have the whole world ahead of you.
Hanh Brown: 01:16:35
And this is just one out of many learnings ahead. Whether it’s life lessons, experiences outside of your school, in your workplace, and then future family life. This is one out of many learnings ahead and I would embrace it, but proceed with caution.
Will Trevor: 01:16:54
Thanks, Hanh. Lori?
Lori Mazor: 01:16:58
Yes. So I, I pivoted my career to make this what I do for a living. Um, so I am a little biased in this way. I train people to learn generative AI and I train them in the same way that you would learn how to play the guitar, um, or, you know, play a musical instrument, um, or learn how to paint, which is, uh, it’s very hands-on.
Lori Mazor: 01:17:20
I don’t think this is an intellectual pursuit, and I don’t think it’s one that we are going to do by watching online. Um, uh, videos. I, I think this is one we do in community. Um, and actually I think the greatest opportunity, uh, post pandemic is for us to join learning communities and use the tools and pick one, uh, ignore the, the overload of what I call AI junk food.
Lori Mazor: 01:17:47
Pick one foundational model, whether it’s ChatGPT or Bard, or if you want an image generator, it really doesn’t matter. What you’re learning is exactly what I’ve been talking about with this balance between control and chaos. You’re learning how to communicate effectively with a machine, and you’re also seeing its flaws, and I think that is what’s worth
Lori Mazor: 01:18:08
having a conversation about, right? So I think pick one experiment, do it in community, get yourself trained and educated, and then we can have a more meaningful conversation at an academic level about what this means.
Will Trevor: 01:18:23
Thanks, Lori. James. Re-skilling, re-retooling? How do, how do we do it?
James Intriligator: 01:18:30
Uh, I, I think that it, again, it’s something that needs to be done throughout, sort of, throughout the educational system. It, it should be that we are teaching kindergartners how to ask better questions, how to understand how to frame a question so that someone answers. I mean, a lot of these are skills that are light skills in any case, not just
James Intriligator: 01:18:48
ones relevant to ChatGPT, and LLMs, right? Asking, bringing the right question. Lots of quotes about how that’s the most important part is coming up with the right question. In ChatGPT, LLM world, that’s even more important than any anywhere else. Figuring out what you really want to know, mapping out a path to get from where you are, to where you want to go,
James Intriligator: 01:19:05
looking for opportunities to explore the, the broader landscape, et cetera. Those are sort of the skills that we should be teaching. Maybe not kindergarten, but um, at the very latest, I would say middle schoolers should be learning that. And in parallel, learning how to use it responsibly, right? You, uh, how, how to go about,
James Intriligator: 01:19:20
um, using it to generate. Better answers, better responses, better knowledge. Um, thinking about the implications from A DEIJ broader social cultural perspective, all of those are topics that we should be teaching them to engage with and think about, and even using ChatGPT to kind of investigate, you know, I mean, I think that if you’re interested in
James Intriligator: 01:19:39
issues, uh, like, um, DEIJ issues, let’s say social justice issues, et cetera, you can easily ask ChatGPT to start investigating those issues with you. And I’ve found that I, I’ve learned so much by walking through, uh, fields of knowledge, kind of half led half leading. It’s kind of a collaborative exploration with ChatGPT. Um, and, and I think that that is, Kind of
James Intriligator: 01:20:02
the skill and the approach that we need to be, uh, taking in education that we need to teach people how to use it that way. How to use it responsibly. Uh, I’ll just go back to something I had mentioned earlier. So if, let’s say you’re in high school and you have to write a AP English essay and it’s due Friday, et cetera, uh, it’s very easy for kids now to just go to
James Intriligator: 01:20:24
ChatGPT and have it write the essay and, you know, you could ask it to make it a little bit dumbed down so it doesn’t sound like it was written by ChatGPT. You could ask it to include particular themes, you know, that’s pretty good, but make it 10% shorter. Can you make it 20% shorter and sound a little bit more written at a slightly lower, uh, Linguistic
James Intriligator: 01:20:38
sophistication level, you could do all of those things and end up with a wonderful final artifact. But I think both the training and the assessment should actually start to ask to see that path that you took. Uh, and, and part of what you’ll be assessed on is not the final deliverable, but more the, the journey you took to assemble that final deliverable.
James Intriligator: 01:20:58
It’s like more, more like an artist’s assessment, um, a critique session where you wanna, you know, you don’t wanna just see the final industrial designed product. You wanna see all of the sketches that led up to it and all the explorations. Which textures did you explore? Which colors, how did you look at them, et cetera. And I think with Chachi Bt, as you’re
James Intriligator: 01:21:12
sort of partner on that exploration, you can really go very far. But you should also be documenting that path and that’s how maybe you’ll be assessed. But again, I’m not really sure.
Will Trevor: 01:21:25
Thanks, James. Partha?
Partha Ghosh: 01:21:27
Oh, it’s a million dollar question. I see it the two sides to it. One, the how to skill and re-skill people who be serving. I would call the entire AI in universe, and obviously that’s where people have to learn about how to manage ChatGPT. But the other side, I feel it’s not a question of re-skilling, but all of a
Partha Ghosh: 01:21:55
sudden you have got this intelligence with which you can converse, comes into play. It’s like when the electricity came into play, everyone did not learn about electricity and how to make the bulb or how to work in the power stations. But with the light came new ways of thinking. People can work 24/7, they can work in the night.
Partha Ghosh: 01:22:16
Similarly, when the intelligence have come into play, I think their education has to focus on how to use this new power for the betterment of society. And then comes a question, education of whom? Are we talking about the top 10% of the people, the world, or are we talking about the bottom of 80%? I look at the people at the bottom
Partha Ghosh: 01:22:35
of the 80% bottom, 80% people in the small villages in Africa, in India, in Brazil, in Indonesia. Unfortunately, through my Boston pledge activities, I see the intelligence there, except they don’t have the access. So how good education be modified using the power of ChatGPT, so that you can converse in the local language, not in English, but the other thing that
Partha Ghosh: 01:23:03
we have done to the world, we have lost a lot of the great languages because everyone has to learn English to belong ChatGPT gives you the power to speak in your mother tongue. If you are a village in Chile, you speak in the local Chilean language, you talk to someone in Bangladesh, Bengali, but they, the cloud would do the translation for you.
Partha Ghosh: 01:23:24
So that’s the kind of education that we could bring to play where, and you know, we do not have to go to DeVos the top. 5% of the population of the one person that goes. But when people could begin to conference with each other when they’re in the, in the ground, in a village, in a field, and exchange best practices in their own mother
Partha Ghosh: 01:23:45
tongue, that would open up a new world. And that’s the kind of education that we have to bring to play. So question is education for whom? And education for the bottom 80 person. If that could be reformed, reimagine using the power of ChatGPT, then we have done service to the world and that’s why the were one of my comp, the companies I’m involved with in queue is working on.
Partha Ghosh: 01:24:11
So it’s possible it’s happening. And I think that’s where the world could go.
Will Trevor: 01:24:16
Thank you, Partha. Um, so we just have seven minutes remaining and I’d like to pose this question to, to all of you. Uh, Hanh I think it was you, uh, who started us off by saying that, you know, that this is the start of maybe even greater developments in technology, uh, which, which, which kind of reminded me of, uh, of, uh,
Will Trevor: 01:24:35
Douglas Adams’ “Hitchhiker’s Guide to the Galaxy” when they invent the computer, Deep Thought in the first, when they ask it about life, meaning, uh, uh, life life’s, the, the earth meaning and everything. And it says, well, actually, that’s the computer that comes next that I will invent will actually help you, uh, answer that.
Will Trevor: 01:24:53
So Hanh, uh, help us look, uh, look to the high horizon here. What is next? What is coming next? What would, what we, what should we be watching, uh, in, in terms of what the future trends have in store for us?
Hanh Brown: 01:25:06
Well, as you know, some of the technologies that we are seeing, some are still in beta, right? And even when beta it is amazing. For instance, Wolfram plugins in ChatGPT, amazing. Science, technology, engineering, and also Runway ML. Amazing, right? So, what should we look ahead?
Hanh Brown: 01:25:32
Well consider where we are right now with some recent development and even when, uh, it’s only in beta, I think it’s almost like the sky is the limit. So I’m really excited to engage in these tools and technologies, in all regards. I’m learning, Hugging Face, Python, Wolfram, Auto-GPT. So, the key here is keep learning because it’s evolving so fast.
Hanh Brown: 01:26:03
So that, that would be my take on what’s ahead of us, is that we, we, in order to keep up, we really need to be learning every day.
Will Trevor: 01:26:11
Thanks, Hanh. Bhaskar. What’s, uh, what’s on the horizon for us?
Bhaskar Chakravorti: 01:26:15
Yeah. In the spirit of learning, one thing I’ve learned is it’s mug’s game, trying to predict, uh, uh, the future, right? So I, uh, won’t even step in there. Uh, uh, what I would say perhaps is what I would like to see in the future, and, uh, what I would like to see is, uh, generative AI, uh, actually having some byproducts, not
Bhaskar Chakravorti: 01:26:35
necessarily in generating, you know, uh, new visuals or new essays and so on. I think that’s all very, very useful and relevant, uh, uh, uh, material. Um, what I’d like to see is, uh, all the innovations that are going to go into, uh, the creation of future generations of, uh, of GPT uh, and, uh, The work that, uh, very smart people are gonna be involved in the harnessing of data.
Bhaskar Chakravorti: 01:27:04
Uh, uh, the thinking around, um, creating ethical guardrails and, and, and frameworks, uh, uh, not just relating to generative AI, but AI more broadly. I’m hoping that it has a spillover effect in other applications of AI. And I would say, just to kind of go back to, I might be sounding like a broken drum. I think AI’s biggest application
Bhaskar Chakravorti: 01:27:30
isn’t saving human life, and so far it’s done a terrible job, but it’s potential is enormous. So, even before, uh, generative AI hit the headlines in, in the end of 2022, in the mid stage or early parts of 2022, another landmark development happened in AI when Alpha Fold, uh, basically, uh, developed the ability to predict the structure of every protein
Bhaskar Chakravorti: 01:27:56
in, uh, the human body. Now, what that means is it could potentially open the door to massive amounts of revolutions in drug discovery, vaccine discovery, new forms of treatment. So, I’m really hoping that the future is one where the spillover of all the talent and effort and capital that’s going into generative AI can actually help us deliver better outcomes in AI, in healthcare,
Bhaskar Chakravorti: 01:28:20
AI, in saving human life, whether it’s in health or crisis response, or to a variety of other ways in which this kind of technology could be put to use.
Will Trevor: 01:28:32
Thanks, Bhaskar. James, what, what should we be looking at on the horizon?
James Intriligator: 01:28:36
Uh, well, it, it, like, like they all said, it’s very hard to make any predictions. When I first started playing around with ChatGPT back in December, I was thinking, oh man, a year from now they’re gonna have things they can do X, Y, or Z. And within three weeks they had things they can do X, Y, or Z. And I thought, wow, okay.
James Intriligator: 01:28:51
In like six months they’ll have something they could do A, B, C, and then within two weeks, A, B, C was happening. I mean, the speed of acceler, the, the speed of growth here is, is just kind of scary actually. It really is. I mean, uh, I can’t even imagine what things will be like six months from now. The ability to generate video,
James Intriligator: 01:29:05
generate music, I mean, all of these things are happening. Generate agents that generate other agents to take care of complex tasks. You can, uh, spawn sub-agent. I mean, it’s crazy. I, one thing I’ve just been thinking about recently is the idea that maybe this should be thought of as a public utility. Maybe it should be
James Intriligator: 01:29:21
provided free of charge. Maybe the government should just eminent domain, take over open AI and offer it up as a free service available to everyone. That it should be seen like water or air or food. It should be a human right to have access to ChatGPT. Uh, it is incredibly empowering to have that access and to not have
James Intriligator: 01:29:39
it is incredibly disempowering. It puts you at a real disadvantage from a social justice perspective. I think that’s a huge issue. Uh, I’ll stop there.
Will Trevor: 01:29:47
Thanks, James. We’ve got just two minutes left. So, Partha and Lori, if I could have your, your, your headline thoughts on, on what is next?
Partha Ghosh: 01:29:54
Let Lori go first and Partha will finish with whatever time is left.
Lori Mazor: 01:30:03
I will add that, um, the dirty little word, uh, the Metaverse is not going away. And in fact, uh, I think what we are going to see with AI and what we are already seeing in the signals, um, from those things that are in beta, is that synthetic reality will be in our hands in a very short period of time. And the moment at which my 15 year
Lori Mazor: 01:30:24
old daughter can use natural language to envision a space that she wants to invite her friends to come over, um, and using, you know, well whatever version of, um, of AR versus VR, we are, we are at, um, this is the moment that we’re going to move from this post-truth world into, uh, a realm of possibility. And, you know, and the matrix will be a reality.
Lori Mazor: 01:30:53
So we have a lot to, um, to deal with beyond just AI.
Will Trevor: 01:30:57
Thanks, Lori. Partha? Yeah, one minute.
Partha Ghosh: 01:30:59
I think uh, you know, it’s has been all kind of touched on. The world would be different because of ChatGPT, because of AI. They clearly scratching the surface more to come given what more to come as Bhaskar and all of you touched on. We cannot really bet on what that more might be. But what we could definitely bet on
Partha Ghosh: 01:31:18
that the choices and the options in front of us will exponential increase. That’s point number one. Some of them will open up new possibilities. Some of will create problems, that’s for sure. We talked about democracy. We do not know how the democracy would evolve, but democratization
Partha Ghosh: 01:31:40
of knowledge will happen. So that gives me hope that as Bhaskar you were talking about, that we will able to bring a new equation for the development of humanity. That’s very important because fortunately I’ve been exposed to a hundred dollars per capita environment to $200,000 per capita environment. And it is shameful that we can make
Partha Ghosh: 01:32:02
those environment coexist today. We have to solve that problem. So in that spirit, the the point that I would like to make for the younger people, the choices are many life cycle of choices would be shrinking because that’s where we have to be extremely fast. We have to be knowing how to engage with opportunities cause that’s the shrink. And most important would be if AI
Partha Ghosh: 01:32:27
could do the logical work, the, uh, mathematical work, the Ian work, could we unlock through the education system, the imaginative power of society, the imaginative power of individuals, which is the right brain, and in the process, create a new world that each one of us could be proud of. So, I think that’s what I would like to emphasize, that we have
Partha Ghosh: 01:32:56
an opportunity to serve humanity better and we should work with it. And of course, I would not use the word guardrails, but be concerned citizens, to be able to avoid the unintended consequences. You know, the term unintended cons is too frequently used, or we experimented, but unintended, we cannot afford that because anything unintended in
Partha Ghosh: 01:33:19
the world of AI would be extremely dangerous, which we cannot afford. So we have to take a holistic view as to how we engage with the world of ChatGPT, the world of AI, and in the process create a better world that we are proud of.
Will Trevor: 01:33:34
Thank you, Partha. And that just takes us to up to two o’clock. So, uh, thank you very much to all of the panelists for coming and talking to this, uh, very topical, uh, issue. You, you brought a range of different perspectives and it was a, a highly insightful and interesting, um, uh, discussion.
Will Trevor: 01:33:53
It really was. Uh, Bhaskar Chakravorti. Thank you to you. Uh, Hanh Brown. Thank you to you. James Intriligator. Thank you. Partha Ghosh. Uh, and, uh, Lori Mazor. Thank you to all of you for freely,
Will Trevor: 01:34:02
uh, giving of your time, uh, today. Uh, thanks also to those of you who, uh, came attended. Uh, this keynote debate will be taking a break over the summer, but we’ll be back, uh, in the fall term with a different, uh, topic to look at. Uh, we are gonna take notice on the questions. We tried to amalgamate
Will Trevor: 01:34:18
some of those questions. Some of those questions were answered on live, but we may well make a post in which we, uh, answer each of those questions. But, uh, uh, please enjoy the rest of your day, travel safely. Thanks very much for coming today. Goodbye.