Everyone, it seems, is talking about ChatGPT - the generative artificial intelligence application that you can chat to like a very intelligent human and instruct to write things previously done by humans.
In this episode, Aakrit Vaish, Co-Founder & CEO of Haptik, an Indian company that makes AI that interacts with people by voice or text, says the developments of the last few months are a real breakthrough and that we are entering a golden age of AI.
Check out all our podcasts on wef.ch/podcasts:
Join the World Economic Forum Podcast Club
Podcast transcript
This transcript has been generated using speech recognition software and may contain errors. Please check its accuracy against the audio.
Aakrit Vaish, CEO of Haptik: By far the most exciting thing to happen in my industry has been ChatGPT. It's been an amazing revelation of what the technology can do.
Robin Pomeroy, host of Radio Davos: Welcome to Radio Davos, the podcast from the World Economic Forum that looks at the biggest challenges and how we might solve them. This week: everyone's talking about it, but how will AI chat technology change our lives and work?
Aakrit Vaish: There's always a reason to be excited and sceptical at the same time.
Robin Pomeroy: We hear from this company that's been in the business of AI chat for a decade now.
Aakrit Vaish: In the 50s or 60s was when the first AI program was written. And, you know, there was a period which was referred to as the AI winter, during roughly the 70s or the 80s. The next 10 or 20 years is really going to be the golden age of AI.
Robin Pomeroy: Aakrit Vaish, co-founder, CEO of Haptik, an Indian company that makes AI that interacts with people by voice or text, says the developments of the last few months are a real breakthrough, but...
Aakrit Vaish: Today, we are at the peak of inflated expectations, where everybody thinks that this is going to absolutely change every aspect of my life. It will, to a great extent, but I do think there's going to be a period where suddenly a lot of us will realise that, okay, look, it doesn't do this, it doesn't do that, it doesn't do all these things. And that's fine. The world needs to be patient as we go through that cycle.
Robin Pomeroy: Subscribe to Radio Davos wherever you get your podcasts or visit wef.ch/podcasts where you'll also find our sister programmes. Meet the Leader, Agenda Dialogues and the World Economic Forum Book Club Podcast. And join the conversation on the World Economic Forum Podcast Club on Facebook. I'm Robin Pomeroy at the World Economic Forum, and with this look at the brave new world of generative AI...
Aakrit Vaish: Now you're going to see a lot of really, really, really cool innovation happen.
Robin Pomeroy: This is Radio Davos.
Everyone, it seems is talking about ChatGPT - the generative artificial intelligence application that you can chat to like a very intelligent human and instruct to write things previously done by humans. I asked it, for example, to write an introduction to this podcast, and I’ll read you what it wrote.
This is, first, what I told ChatGPT:
"Write an introduction to an episode of the World Economic Forum podcast Radio Davos that discusses the potential applications of generative AI, the risks and potential for jobs, and why this is the start of a golden age of AI. It should last around one minute."
OK then. Within a second or two it wrote this:
"Welcome to Radio Davos, the World Economic Forum's podcast on the future of the global economy. In this episode, we will be discussing the exciting potential of generative AI, its implications for job markets, and why we believe this is just the beginning of a golden age of AI.
"Generative AI, also known as GAI, is a type of artificial intelligence that can create new data from existing data, allowing it to generate realistic images, videos, and even text. This technology has already been used in fields such as gaming and digital media, and its potential applications are vast, ranging from improving medical diagnoses to creating more realistic virtual assistants.
"However, the rise of generative AI also poses risks for job markets, as it has the potential to automate tasks previously performed by humans. This has led to concerns about job displacement and the need for reskilling and upskilling.
"Despite these concerns, we believe that the development of generative AI marks the beginning of a new era of innovation and progress. By harnessing the power of AI, we can unlock unprecedented advances in fields such as healthcare, finance, and transportation, creating new opportunities for growth and development.
"Join us as we explore the potential of generative AI and the implications for the future of work, in what we believe is the start of a golden age of AI."
Not bad - it wrote that in a few seconds, although Radio Davos looks at much more than ‘the future of the global economy’ - but let’s not quibble.
So while armchair editors are having fun like that with ChatGPT and other generative AI systems, what do people who already work in the field think of it?
Aakrit Vaish is the Co-Founder and CEO of Haptik which uses conversational AI to enable companies to interact with their customers via voice and text, such as ordering goods or getting after-sales service. They have been doing it for 10 years, but even with that much history, Vaish recognises we are at the dawn of a golden age of AI which will change pretty much everything we do.
Aakrit Vaish spoke with my colleague Kateryna Gordiychuk.
Kateryna Gordiychuk, Head of Social Video Production, World Economic Forum: Can you explain what is conversational AI as we know it today?
Aakrit Vaish: When I started the company in 2013, we're going to finish ten years in August, I remember talking about what we are doing and you know, at that time we used to call it 'messaging' and sometimes 'chat', sometimes 'chat robot'. And it took a long time to explain to people, you know, what we were trying to do.
And it's really funny what's happening now over these last few weeks and months where all of a sudden nobody has that question anymore because of what's happened with ChatGPT and Open AI has done.
So in a nutshell, what what's conversational AI? It uses what we call natural language processing or now today they are referred to as large language models where we look at various data sets that are usually in the form of some form of text that is science that runs over that set of text information, which is basically referred to as data science. You understand the meaning behind what's going on in that data, train what you call, let's say, a model to then give answers to questions that people can ask in natural form.
Today, a lot of this may sound, okay, that makes sense, but, you know, eight or ten years back it didn't. Fundamentally, conversational AI is a technology where you can converse with any AI engine for topics and questions ranging from what's the weather to customer support within the bank or telco or retailer to even, you know, personal concierge services like your own assistant.
Kateryna Gordiychuk: It is true that it's accelerated so much and it's changed so much in the past years, and now we know what it is. We don't really need an explanation. Most people have interacted with this type of AI at some point in their lives, but why does it have to be 'conversational'?
Aakrit Vaish: So, you know, when we started also, we continue continued to believe that conversation is the most natural form of getting something done. Right. So, like, for example, you and I are having this conversation or in a day we have various forms of conversation with lots of people about lots of things ranging from personal to work, professional and so on and so forth. Humans are designed to have conversations and get things done.
Taking that to the next sort of level, if you want to now get things done in a faster manner, you know, it's really the most natural form to also engage with some sort of an AI program.
Kateryna Gordiychuk: Of course, right now with with ChatGPT and other discussions, there's still a lot of people who never use AI in their day to day. It's probably going to change really fast, but some people still seem to be quite afraid of of the change it might bring. As somebody who has been working on this for a decade, why should we learn from conversational AI or work with it as opposed to being afraid of it?
Aakrit Vaish: AI by the way, is a field that has been around for, let's say, roughly 60 years now, or maybe 70 years. In the fifties or sixties was when the first AI program was written. And, you know, there was a period which was referred to as the AI winter during roughly the seventies or the eighties.
I think now from here on forward is going to be whatever the opposite of the AI winter, which is the next ten or 20 years, is really going to be the golden age of AI. And just like the golden age of any technology, whether it was the internet by itself, whether it was the smart phone or whether it was social networking, there's always a reason to be excited and sceptical at the same time. Any new technology innovation causes perhaps 80% excitement and 20% nervousness.
Now with specifically this technology, because it feels so natural, because it's conversational. I think people see the scepticism a little bit more obviously and a little bit more upfront.
Like, for example, when the smartphone came out and you gave you a location permission for using various apps, most people did not have the intellect or the knowhow to imagine how their location data may get used. But when you're chatting within the AI, everybody can imagine that, oh my God, like, is this going to replace my job?
But I think that's the only difference. In reality, I'm always an optimist that with every new technology innovation, net-net, we will move the world forward. We will move jobs forward. We will move the economy forward. And everybody will learn to live in this new AI-first world.
Kateryna Gordiychuk: You're saying that it's just us not knowing how to use it yet or not being fully familiar with it that makes us feel a bit threatened. So I guess the advice is to try to get to know it a bit better or what it could do for us in our jobs, in our daily lives?
Aakrit Vaish: So look, it's new, right? So I think everybody is going to figured out how to pick it up and learn. You know what I would recommend? Just like I recommend just like every other technology wave. Again, when the internet came out of the smartphone came out, you didn't necessarily think so much every day, how am I going to make the best use of the smartphone? Entrepreneurs and developers made that possible for you. The founders of Google created Google to make the internet a lot more accessible and easy for you to use. Similarly, the founders of Facebook and Instagram and WhatsApp created those apps for the smartphone, and then you just found amazing utility. So similarly, let the AI developers of today and tomorrow create those programs, apps and companies that will make it much easier for everybody to use it for various use cases. And along with that, I would encourage every budding businessperson, entrepreneur or software engineer to give their own shot at potentially building some of these AI apps and companies, because the use cases and possibilities are limitless.
Kateryna Gordiychuk: It's very interesting that you said that you think that there is going to be in the next decade a golden age of AI. What kind of industries has been used in the most, and I know it's probably a long list of industries, in the past years, and where do you predict it might shift in the future years?
Aakrit Vaish: When it comes to conversational AI, the largest application so far has been broadly within the customer support use case. So I'm sure you or anyone of us have at some point encountered a support agent that is AI driven either on a phone call or the voice, or even in, let's say, text and chat bots. That, I would say, has been the largest application of conversational AI that has been low hanging fruit, that has been things that are fairly obvious.
Now you're going to see a lot of really, really, really cool innovation happen, honestly, across use cases, across industries. I don't think it's going to be limited to anything, any one thing.
For me, I particularly feel excited about just this entire space of content creation, whether you're creating long-form blog posts or posts or images or video, podcasts, I think a lot of these things are going to become so much simpler to do with advanced large language model capabilities like some of the stuff we're seeing.
But I think this could transcend to lots of critical industries. You know, like, say, for example, healthcare, right? What if an X-ray could just spit out the results versus the doctor kind of having to sift through it on their own? That's a very simple use case. So just tons and tons of applications that could happen in such a critical industry. You know, education, the way we think about it. The tools that are going to be built for teachers and professors to assist them in a way that they can make the entire experience a lot more engaging and productive. But again, be sort of phenomenal. So I think that we're going to see applications that I don't even think we can imagine today.
Kateryna Gordiychuk: When you were saying about the most common application and what comes to mind using conversational AI and this is in customer service, I was thinking, you know, calling the bank and trying to get a customer service representative to talk to me, but instead I would first talk to conversational AI, and I'm sure everybody has that experience has had one at least where they feel frustrated because it's not yet at the same level as a person would help us. How do we bridge, how do we make sure that some things that a human can do, there's still a person for that in that profession or in that role. And then there's also space for conversational AI?
Aakrit Vaish: Absolutely. So look, that's what I mean when I say that every technology moves the industry forward. I get this question asked, I don't know, like ten times a week, which is that, look, is conversational AI going to take away jobs? And my consistent answer is no.
And the best analogy I will give to you is banks and ATMs. Prior to the technology of ATMs, the most common reason why people would walk into a bank would be to take out money. So imagine there's no ATMs and 90% of the reason why somebody goes inside the bank is to say, I need to take out some cash, to take out some money. As a result, this is, you and I don't even remember these days, but banks used to be crowded, there used to be long lines and so on. Here comes the ATM, which automates really what should be a very simple task. You know, you go down to a machine, you put your details and now you get the money out. And then the ATM gradually started doing more and more things. Now, today, when you need to go to a bank for more complex stuff like change of address, close my account, I need a loan. It's such a beautiful experience, right? There's no lines. The agents or the bank tellers are welcoming. They have time for you. They can actually solve your core problems really, really well.
Take that analogy toward every single use case, which is that those jobs are not going to go away, but rather those jobs are going to get more advanced. Those people are going to learn to do much more complex tasks which will fundamentally improve all our lives, including creating better jobs and skills for all these people.
In addition to that, I also believe, and this has been said many times, I think many polls have said, this is one of the most popular jobs in the next sort of eight or ten years is going to be that of any trainer, which is somebody who can, you know, look at all this data and ensure that they put the guardrails behind these engines so they can spawn activity.
Kateryna Gordiychuk: Of course, you've been in the field for so long. What do you see, what's the most exciting change for you at your own company, at Haptik? And how is this development with ChatGPT helping you or hindering you, or how is it changing your day to day? What are you looking forward to?
Aakrit Vaish: Somebody told me a few weeks back, we always get the idea right but the decades wrong. That's kind of what's happening with us, which is that by far the most exciting thing to happen in the industry has been ChatGPT in all these ten years. It's been an amazing revelation of what the technology can do and how it can really advance the capabilities of conversational AI. You know, we are still obviously trying to get our heads and arms around it in terms of what's the best way to use it for us.
But absolutely, it's complimentary to our product. We are already working across a few different APIs of ChatGPT and OpenAI, to integrate them within our solution. And our job, our core business, is essentially to use conversational AI to make really great solutions for brands and businesses. So we are now going to leverage a lot of the core charging tech to add to our existing technology to improve these solutions for our customers.
And AI, and conversational AI, is particularly a candidate for this happening, which is if you've seen kind of the hype cycle of innovation, what tends to happen is the moment the technology gets launched, people get very excited about it. So you reach like this point we call the peak of inflated expectations and then it doesn't work. So then you go back down and you say, okay, listen, this is rubbish, it's not going to work. And then eventually you'll find the right set of use cases for it. And then you start going on what we call the slope of enlightenment, which is then you start saying, okay, look, here is the thing that really, really works well for.
So all I'll caution is that I think today we are at the peak of inflated expectations now where everybody thinks that this is going to absolutely change every aspect of my life. It will to a great extent. But I do think there's going to be a period where suddenly a lot of us will realise that, okay, look, it doesn't do this, it doesn't do that, it doesn't do all these things. And that's fine, you know, but very quickly we'll be on that steady state slope where we'll be able to see all productivity and gains come out of it. So I would just say that the world needs to be patient as we go through that cycle.
Robin Pomeroy: Aakrit Vaish is the Co-Founder & CEO of Haptik.
For another great radio Davos episode on AI, check out our interview with Professor Stuart Russell - one of the world’s leading experts on the subject. It’s called ‘The promises and perils of AI’ - find that in our back catalogue on your podcast app, or at wef.ch/podcasts, where you will also find our sister podcasts, Meet the Leader, Agenda Dialogues and the World Economic Forum Book Club Podcast.
And if you are looking for other podcasts on AI, check out In AI We Trust, co-hosted by the Forum’s Head of AI Kay Firth-Butterfield
Please subscribe to Radio Davos wherever you get your podcasts and please leave us a rating and a review. And join the conversation on the World Economic Forum Podcast club on Facebook.
This episode of Radio Davos was written and presented by me, Robin Pomeroy, with reporting by Kateryna Gordiychuk. Studio production was by Taz Kelleher.
We will be back next week, but for now thanks to you for listening and goodbye.
Podcast Editor, World Economic Forum
Video Producer, World Economic Forum
Devanand Ramiah
December 6, 2024