Again, a very warm welcome to all of you present here and all our viewers watching at home. This is the AI Summit, and with an organization like the ATA Group, we've always decided that in every new evolving world feature or aspect, we have to be the leaders. And so is India thinking that we have to go straight forward, straight ahead in AI. Big dreams, and the AI dream that we are going to talk about — interestingly, I want all of you to clap for — but this is the second Indian that we have on stage.
Isn't this amazing? Like in our AI Summit, we have these important companies. Your company, Rubrik, is based in America, and he's the founder and CEO of Rubrik, and he's an Indian. And interestingly, he's from Madhepura, Bihar.
Real Bihar, not Paschim Bihar.
Wow. Interesting. A huge round of applause for this as well, because from these little towns of India, people are emerging as the leaders in AI. And your company, Rubrik, essentially works on data security and cyber security and AI. From what I have gathered about your organization, I want to begin from there. How important is it that we have a disciplined execution of everything that we are talking about today in the field of AI? How do you see that happening?
Anjana, first of all, very excited to be here. Look, AI is 100 times more opportunities and 100 times more risk because models hallucinate and also they are prone to compromise. And if they are compromised, they can potentially do 10 times more damage in one-tenth of the time because they can write code, reimburse your customers, and run your business processes. And somebody sitting in North Korea can control your whole business. So every business has to think about how do they deploy AI with confidence and trust.
Yeah. But the way you've mentioned North Korea, thanks for drilling that in. This is the scare factor. You keep saying that 100x opportunity, 100x risk, but people are scared of the risk. Like, I'll tell you from my example: when my AI avatar was being made, and Kalpuri, our vice chairperson, decided that we should do this, I was a little scared, to be honest, because I was scared that — I mean, she's going to be doing anchoring on my behalf, what will happen to me? But then eventually she lured me by saying that, imagine you can go on a vacation and she's going to anchor for you.
[Laughter] You can be with your daughter at Oxford, and she's going to anchor for you. And then I said, yes, ma'am. So, but that's great. But then from here, it has seeped into our daily lives, and closing the eyes will not push AI away. It's here. How do we build this bond with it? How do we make them the force multiplier for my performance?
For successful AI, it is very important to have the right guardrails. So guardrail means that what AI can do and cannot do. For example, if there is a Rubrik AI that is allowed to send email, I don't want that AI to send email to everybody in my company on every topic. So even if they have permission, there has to be a guardrail, and that guardrail has to be defined and enforced real time. And that is the challenge of AI, and this is what Rubrik is focused on. What we are telling our customers is: deploy AI with confidence. Because if you look at countries such as India, AI has tremendous promise to deliver healthcare, education to far corners of India. But if people don't have digital trust on what they are interacting with, this whole exercise is moot.
AI cannot replace me because I live in the future and AI lives in the past.
Okay. So here is the basic theme that data is the engine for AI, and you work on data security. How does that whole process go, and how important is this?
Data is the key ingredient of AI. If there is no data, or if you don't have confidence in your data, then garbage in, garbage out, and the whole AI stack is of no use. So what we are saying is that we will make the right data available to the right user on the right platform at the right time to really deploy AI. And if you think about it, making the data available is the key factor. And so what we are saying to our customers is that deploy AI, take your data, make your model completely optimized to your own environment so that models don't hallucinate and you have high trust and confidence.
Models don't hallucinate. I want to talk of the doomsday scenario that crosses a lot of our minds. What if there's a massive cyber attack? Like there's an attack and the wrong people take control of it. What happens then to humankind?
See, cyber attacks are inevitable, just like death and taxes. And for far too long, we have been focused on prevention and detection and not thinking about how do we come back up after an attack. Exactly. So the whole thesis of us starting Rubrik was that how do we ensure that you can successfully recover and keep your services up and running? Because if you are running a hospital, you can't wait three weeks for cyber recovery to happen. You have to bring the hospitals back up because you have to deliver healthcare to all the patients. So every business, every government organization has to be prepared with resiliency, and that resiliency means: do you understand the risk, and do you have the ability to recover?
We have with us a minister from Canada. I think he's out of the room right now. Oh yeah, yeah. But interestingly, you know, I showed him my AI avatar, and he also said, 'I want someone to replace me now. Where's my AI avatar? I want to sign up for this now. I've got to have my vacation in my hometown, and I've got to do this.' But coming back to that point, people have everything on their smartphones, right? From banking to everything. How does that make them vulnerable to any kind of attack, as you call it? And because you're a data security company, a cyber security organization, how does one prepare for that? How vulnerable does it make us in terms of our money, everything?
Look, if you look at the last like 10, 15 years, the digital intensity of our lives has increased because we get up in the morning and look at the maps — that's how we drive to work. We interact with each other over emails, not calling and talking to people all the time. So as the digital intensity has increased, so has the cyber risk.
And so everyone has to now think: am I really prepared? I'm doing the basic hygiene. Do I have multifactor authentication? Do I trust this person that I'm giving information to? So that's the basics. And then you need to have the ability to have been able to recover from a bad incident. So everybody has to think about that.
What exactly do you mean when you say we have to be — like we had other guests also who were talking about becoming AI aware? What exactly is this AI aware business? I mean, downloading some apps, navigating through it, or how does one go about it? Like so many students sitting here, how can one become AI aware?
AI is all about: can we do knowledge work automatically without human intervention? And so if you think about when I'm traveling to India, my assistant has to find a hotel, airplane, local transportation. These are knowledge work. So if you fast forward with AI deployment, all of it — she's not trying to reinvent the future, she's just taking the available information and compute what is the best way to travel. That AI can do. That is the knowledge work. So all of us have to think: what is our repetitive task, and can AI do that repetitive task and make local decisions without our intervention? But AI can only work on the past, and that's why I'm saying that us humans have to work on intuition, where we connect new dots, create new things. So all the jobs, all the opportunities around connecting new dots is going to be human work, and everything which is coded knowledge and data on which you can run the work will be AI.
You've traveled from Madhepura to California, built this company of yours, and what is your basic background, if I may ask?
I'm an engineer turned venture capitalist turned entrepreneur.
Wow, that's quite a journey, very interesting. But I get messages from my relatives who tell me that you've just been talking about this diabetes medicine and is it really good? And I say I have not done any such thing. I'm coming to the issue of deepfakes. How does one deal with that? I can't even tell my own avatar's voice from mine. How does one deal with deepfake?
Look, AI-based attacks have gone beyond human comprehension. And we need to go back in time and think how did we solve some of these problems. If you go back to the late 90s, early 2000s, nobody was ready to swipe the credit card on the internet because everyone was worried about: is this a real website or is it a fake website? And then we invented VeriSign that certified individual websites. So you will see innovations around deepfake where you will have a certified AI persona interacting on digital media.
Okay. And that certification is important for individuals.
Absolutely. Because otherwise, how will I know that I'm interacting with Anjana digitally versus somebody impersonating Anjana from a remote country?
Interestingly, I was at the summit yesterday and there were people who were coming up and asking me, 'Are you Anjana or the AI Anjana?' But how does India go ahead? See, it's important that we are not just a market for AI, because this is the challenge we face in most fields. India is seen as the biggest bazaar, but how do we become the drivers of this? How do you think — what are the three or four or five steps that you would suggest for this?
