(upbeat music) >> Okay, welcome back everyone to theCube's coverage of AWS re:Mars here in Las Vegas. I'm John Furrier, host of theCube. Got Alexey Surkov, Partner at Deloitte joining me today. We're going to talk about AI biased AI trust, trust in the AI for the, to save the planet to save us from the technology. Alexey thanks for coming on. >> Thank you for having me. >> So you had a line before you came on camera that describe the show, and I want you to say it if you don't mind because it was the best line that for me, at least from my generation. >> Alexey: Sure. >> That describes the show and then your role at Deloitte in it. >> Alexey: Sure. Listen, I mean, I, you know, it may sound a little corny, but to me, like I look at this entire show, at this whole building really, and like everybody here is trying to build a better Skynet, you know, better, faster, stronger, more potent, you know, and it's like, we are the only ones, like we're in this corner of like Deloitte trustworthy AI. We're trying to make sure that it doesn't take over the world. So that's, you know, that's the gist of it. How do you make sure that AI serves the good and not evil? How do you make sure that it doesn't have the risk? It doesn't, you know, it's well controlled that it does what we're, what we're asking it to do. >> And of course for all the young folks out there the Terminator is the movie and it's highly referenced in the nerd circles Skynet's evil and helps humanity goes away and lives underground and fights for justice and I think wins at the end. The Terminate three, I don't, I can't remember what happened there, but anyway. >> Alexey: I thought the good guys win, but, you know, that's. >> I think they do win at the end. >> Maybe. >> So that brings up the whole point because what we're seeing here is a lot of futuristic positive messages. I mean, three areas solve a lot of problems in the daily lives. You know, machine learning day to day hard problems. Then you have this new kind of economy emerging, you know, machine learning, driving new economic models, new industrial capabilities. And then you have this whole space save the world vibe, you know, like we discover the moon, new water sources maybe save climate change. So very positive future vibe here at re:Mars. >> Alexey: Absolutely. Yeah, and it was really exciting just watching, you know, watching the speakers talk about the future, and conquering space, and mining on the moon like it's happening already. It's really exciting and amazing. Yeah. >> Let's talk about what you guys are working at Deloitte because I think it's fascinating. You starting to see the digital transformation get to the edge. And when I say edge, I mean back office is done with cloud and you still have the old, you know, stuff that the old models that peoples will use, but now new innovative things are happening. Pushing software out there that's driving you with the FinTech, these verticals, and the trust is a huge factor. Not only do the consumers have a trust issues, who owns my data, there's also trust in the actual algorithms. >> Exactly. >> You guys are in the middle of this. What's your advice to clients, 'cause they want to push the envelope hard be cutting edge, >> Alexey: Right. >> But they don't want to pull back and get caught with their, you know, data out there that might been a misfire or hack. >> Absolutely. Well, I mean the simple truth is that, you know, with great power comes great responsibility, right? So AI brings a lot of promise, but there are a lot of risks, you know. You want to make sure that it's fair, that it's not biased. You want to make sure that it's explainable, that you can figure out and tell others what it's doing. You might want to make sure that it's well controlled, that it's responsible, that it's robust, that, you know, if somebody feeds it bad data, it doesn't produce results that don't make sense. If somebody's trying to provoke it, to do something wrong, that it's robust to those types of interactions. You want to make sure that it preserves privacy. You know, you want to make sure that it's secure, that nobody can hack into it. And so all of those risks are somewhat new. Not all of them are entirely new. As you said, the concept of model risk management has existed for many years. We want to make sure that each black box does what it's supposed to do. Just AI machine learning just raises it to the next level. And we're just trying to keep up with that and make sure that we develop processes, you know, controls that we look at technology that can orchestrate all this de-risking of transition to AI. >> Deloitte's a big firm. You guys saw you in the US open sponsorship was all over the TV. So that you're here at re:Mars show that's all about building up this next infrastructure in space and machine learning, what's the role you have with AWS and this re:Mars. And what's that in context of your overall relationship to the cloud players? >> Alexey: Well, we are, we're one of the largest strategic alliances for AWS, and AWS is one of the largest ones for Deloitte. We do a ton of work with AWS related to cloud, related to AI machine learning, a lot of these new areas. We did a presentation here just the other day on conversational AI, really cutting edge stuff. So we do all of that. So in some ways we participate in that part of the, the part of the room that I mentioned that is trying to kind of push the envelope and get the new technologies out there, but at the same time, Deloitte is a brand that carries a lot of, you know, history of trust, and responsibility, and controls, and compliance, and all of that comes, >> John: You get a lot of clients. I mean, you have big names. Get a lot of big name enterprises >> Right. >> That relied on you. >> Right, and so >> They rely on you now. >> Exactly, yeah. And so, it is natural for us to be in the marketplace, not only with the message of, you know, let's get to the better mouse trap in AI and machine learning, but also let's make sure that it's safe, and secure, and robust, and reliable, and trustworthy at the end of the day. And so, so this trustworthy message is intertwined with everything that we do in AI. We encourage companies to consider trustworthiness from the start. >> Yeah. >> It shouldn't be an afterthought, you know. Like I always say, you know, if you have deployed a bot and it's been deciding whether to issue loans to people, you don't want to find out that it was like, you know, biased against a certain type of (indistinct) >> I can just see in the boardroom, the bot went rogue. >> Right, yeah. >> Through all those loans you know. >> And you don't want to find out about it like six months later, right? That's too late, right? So you want to build in these controls from the beginning, right? You want to make sure that, you know, you are encouraging innovation, you're not stifling any development, and allowing your- >> There's a lot of security challenges too. I mean, it's like, this is the digital transformation sweet spot you're in right now. So I have to ask you, what's the use case, obviously call center's obvious, and bots, and having, you know, self-service capabilities. Where is the customers at right now on psychology and their appetite to push the envelope? And what do you guys see as areas that are most important for your customers to pay attention to? And then where do you guys ultimately deliver the value? >> Sure. Well, our clients are, I think, are aware of the risks of AI. They are not, that's not the first thing that they're thinking about for the most part. So when we come to them with this message they listen, they're very interested. And a lot of them have begun this journey of putting in kind of governance, compliance, controls, to make sure that as they are proceeding down this path of building out AI, that they're doing it responsibly. So it is in a nascent stage. >> John: What defines responsibility? >> Well, you want to, okay, so responsibility is really having governance. Like you have a, you build a robot dog, right? So, but you want to make sure that it has a leash, right? That it doesn't hurt anybody, right? That you have processes in place that at the end of the day, humans are in control, right? I don't want to go back to the Skynet analogy, right? >> John: Yeah. >> But humans should always be in control. There should always be somebody responsible for the functioning of the algorithm that can throw the switch at the right time, that can tweak it at the right time, that can make sure that you nudge it in the right direction that at no point should somebody be able to say, oh, well, it's not my fault. The algorithm did it, and that's why we're in the papers today, right? So that's the piece that's really complex, and what we try to do for our clients as Deloitte always does is kind of demystify that, right? >> John: Yeah. >> So what does it actually mean from a procedures, policies, >> John: Yeah, I mean, I think, >> Tools, technology, people. >> John: Yeah, I mean, this is like the classic operationalizing a new technology, managing it, making sure it doesn't get out of control if you will. >> Alexey: Exactly. >> Stay on the leash if you will. >> Alexey: Exactly. Yeah. And I guess one piece that I always like to mention is that, it's not to put breaks on these new technologies, right? It's not to try to kind of slow people down in developing new things. I actually think that making AI trustworthy is enabling the development of these technologies, right? The way to think about it is that, we have, you know, seat belts, and abs brakes, and, you know, airbags today. And those are all things that didn't exist like 100 years ago, but our cars go a lot faster, and we're a lot safer driving them. So, you know, when people say, oh, I hate seatbelts, you know, you're like, okay, yes, but first of all, there are some safety technologies that you don't even notice, which is how a lot of AI controls work. They blend into the background. And more importantly, the idea is for you to go faster, not slower. And that's what we're trying to enable our clients to do. >> Well, Alexey, great to have you on theCube. We love Deloitte come on to share their expertise. Final question for you is, where do you see this show going? Where do you guys, obviously you here, you're participating, you got a big booth here, where's this going? And what's next, where's the next dots that connect? Share your vision for this show, and kind of how it, or the ecosystem, and this ecosystem, and where you're going to intersect that? >> Wow. I mean, this show is already kind of pushing the boundaries. You know, we're talking about machine learning, artificial intelligence, you know, robotics, space. You know, I guess next thing I think, you know, we'll be probably spending a lot of time in the metaverse, right? So I can see like next time we come here, you know, half of us are wearing VR headsets and walking around and in meta worlds, but, you know, it's been an exciting adventure and, you know I'm really excited to partner and spend, you know spend time with AWS folks, and everybody here because they're really pushing the envelope on the future, and I look forward to next year >> The show is small, so it feels very intimate, which is actually a good feeling. And I think the other thing in metaverse I heard that too. I heard quantum. I said next, I heard, I've heard both those next year quantum and metaverse. >> Okay. >> Well, why not? >> Why not? Exactly, yeah. >> Thanks for coming on theCube. Appreciate it. >> Thank you. >> All right. It's theCube coverage here on the ground. Very casual Cube. Two days of live coverage. It's not as hot and and heavy as re:Invent, but it's a great show bringing all the best smart people together, really figure out the future, you know, solving problems day to day problems, and setting the new economy, the new industrial economy. And of course, a lot of the world problems are going to be helped and solved, very positive message space among other things here at re:Mars. I'm John furrier. Stay with us for more coverage after this short break. (upbeat music)