Clips
News
More from HPE Discover 2024

Kirk Bresniker

Hewlett Packard Labs Chief Architect, HPE Fellow & VP

HPE

Harnessing AI for social good: How HPE is championing human rights and ethical innovation

Since artificial intelligence is presenting new cases, AI for social good is gaining steam based on the need to safeguard human rights for enhanced fairness and equality. Hewlett Packard Enterprise Co. has taken the initiative by using AI to uncover hidden signals of enslaved labor and human trafficking. This marks a significant step in addressing these critical issues, according to Kirk Bresniker (pictured), Hewlett Packard Labs chief architect, fellow and vice president at HPE. HPE’s Kirk Bresniker talks to theCUBE about how AI for social good is a game-changer

play_circle_outline Responsible ethical AI principles

play_circle_outline Challenges in governance and compliance

play_circle_outline AI's potential in societal benefit

play_circle_outline AI's role in human rights, human trafficking

Info

Kirk Bresniker returns to theCUBE to dig deeper into the responsible design and use of AI

Kirk Bresniker

Hewlett Packard Labs Chief Architect, HPE Fellow & VP HPE

Rebecca Knight

>> Good morning everyone, and welcome back to theCUBE's Live coverage of HPE Discover 2024. It is Juneteenth 2024. I'm your host, Rebecca Knight, sitting alongside my co-host, Dave Vellante. What a show last night.

>> Wow. I tell you, I needed the eyedrops. Willie Mays.

Rebecca Knight

>> Willie Mays, yes-

>> Say, "Hey, kid."...

Rebecca Knight

>> a legend.

>> Four decades the guy played baseball, amazing.

Rebecca Knight

>> Yes.

>> I never got to see him play, but my dad did.

Rebecca Knight

>> Oh, that's great. That's great. Well, we're going to be talking about a lot of different topics today with our next guest. He is Kirk Bresniker, a good friend of theCUBE, Hewlett Packard Labs', chief architect, HPE fellow, and VP at HPE. Welcome back to the .

Kirk Bresniker

>> Thanks guys. Great to be here, thank you.

>> Good to see you.

Rebecca Knight

>> Yeah. Yeah, so responsible ethical AI, that is something that we're talking a lot about here at HPE Discover, but it's something that actually you've been talking about for many, many years.

Kirk Bresniker

>> I did my CUBE deep dive, and we were talking about these things back in 2017 in Las Vegas and Madrid, and yeah, it was really 2019, and it's great to talk about this today on Juneteenth because human rights was the catalyst for us. We did... 2019, our Global Ethics and Compliance Office did an external audit of Hewlett Packard Enterprise's entire human rights position, and they gave us some great kudos for our labor supply chain and our material supply chain. But they said, "The next three things that Hewlett Packard Enterprise has to worry about in human rights is AI, AI and AI." AI in our products, AI in our processes, and AI when partners build on top of our goods and services to deliver AI outcomes for enterprise. And so the Ethics and Compliance Office gave us a little call at Hewlett Packard Labs and said, "We know how to run governance at HPE. What we don't know about is AI." So we partnered, and from that we came up with five AI ethical principles. I got to lead that team. Originally, I gave us six weeks. It took us about a year to actually come up with all those principles. We started with our purpose to advance the way that people live and work, and then privacy-focused and secured, inclusive, human-focus, robust and responsible. The first three are the human rights focus principles. The last two are the engineering focus principles, because that's what we are. We're engineers who want to have that focus on humanity.

>> And Antonio mentioned those five in his keynote. I want to ask you, whenever I get technical people on, I like to go back and say, "Okay, the AI awakening in November, 2022, people that were inside the technology inside the labs were well aware of LLMs before that." So at what point... We read the Google paper and then all of a sudden the light bulbs started to go off and started to apply CUDA in new ways. And so at what point did you sort of become aware of this notion of large language models and generative AI? And then the second part of the question is, how since the world's awakening has your activity changed?

Kirk Bresniker

>> We were predicting the model. We were expecting this, always expecting this tide to turn. And that's why we started the conversations we did now more than almost 10 years ago. But it was a tectonic shift talking about Willie growing up in the Bay Area, for me, it's always been a fantastic metaphor because I was in fourth grade and our science teacher was talking about the plates going slowly, and then suddenly, boom, it slipped, and then everything changes. So for me, that was that event. November 22, the world changed because it was suddenly so accessible. Everyone across Hewlett Packard Enterprise suddenly realized this is a tool for me. And that made us need to understand how we will drive these technologies faster. We were... January '22, we introduced at the World Economic Forum at Davos a toolkit. So if you were the one person in your C-suite who thought AI was material to enterprise outcomes, how could you have a conversation with your peers and your board and not sound like a conspiracy theorist? Right? And suddenly, here we are 10 months later, and we go into Davos, this year AI was on every surface, everything. And we actually had the AI House at Davos this year opened at 7:00 A.M., standing room only until we shooed them out at 10 P.M. every single day that week. Everyone wants to have this conversation, "How do I use AI? How do I use it responsibly, safely, securely, compliantly? How do I actually enable the enterprise to take advantage of this technology?"

Rebecca Knight

>> Well, that's actually where I was going to go with this conversation because we are all talking about AI, we're talking about it on theCUBE. There's lots of industry conversations. We're talking about it with our children at the dinner table. And what are the world's leaders, the world's decision makers, what are they saying? We have Warren Buffett on one hand who says it scares the hell out of him. And we have obviously other luminaries who are really very bullish.

Kirk Bresniker

>> I think part of the conversation for... And this is also taking off my Hewlett Packard Labs hat and putting on my working group for AI, responsible AI ethics at HPE, the conversation we're having with everyone who comes in front of us is the one about fitness for enterprise use. And so we want to ask about how much energy is in this model? Where'd you get the data? How are you going to give me that understanding of how the decision was reached? And so right now, I think we're seeing a lot of enthusiasm, a lot of real hard sell pitches. And I expect that between innovation, litigation and understanding, we will see the market figure out how can an enterprise, how can a publicly-traded Fortune 100 global enterprise use this? We'll see these things continue to evolve. But right now, the question isn't just, "I know you can do this amazing thing. Can you do it in an enterprise setting? Is this ready for me to adopt into the business process of a publicly-traded enterprise?"

>> So when you think of amazing labs, Xerox PARC, Bell Labs, Hewlett Packard labs, you think about silicon and magic and stuff like that. So when you're talking about governance, connect the dots with the tech, and I think of data governance, but I think more compliance officers, I think backroom function, connect the dots with the technology component and the contributions that R&D makes to this.

Kirk Bresniker

>> So it's been a fascinating series of discussions. We started with the principles. It's one thing to claim you have principles. It's another thing to have them evidence in how you run your business. And that's really been the conversation from after we took that year, got our principles down. Since that point, we've been learning how to live up to those aspirational goals. And all of that has been a conversation. In the governance we're not in the business of saying yes or no. What we're there to say is, "Have you considered?" And it's been really interesting to have an engineering team come in front of us and they'll say, "Here's this fantastic idea." And then you say, "Did you think about this?" And their eyes get big. And they said, "No, I hadn't thought about that." So the important thing about governance is awareness of all the repercussions, all of our principles, both the engineering ones that they're probably well versed in, but then the human rights ones that they may not have considered. And so that's really been the engagement we've had with them; that makes them better engineers, because engineers love a problem. They love constraints. They love boundary conditions. And throwing a new one at them that they hadn't considered, that just gives them more enticement to create a better outcome. But until they know it, they're not going to be able to understand. So that conversation about what is important, what our values, what our principles are, is what helps them to create those great outcomes.

>> So you started with first principles and then what Kirk, then it became just modeling the permutations. Did it become a math problem, a social science problem?

Kirk Bresniker

>> All of the above. And that is the really interesting thing, especially when we think about the repercussions. Part of our responsible principle... And this is back in 2019, and one of my... I said, "We've got to have this one in, part of the principle is you need to understand what resources you're consuming to achieve your AI outcome." Back in 2019, it was like, "Oh yeah, I'm sure it's going to use some power." And now we're like, "We're worried." Current course and trend, we'll use all the power to train one model one time. So understanding both the energy implications, the social implications, that's what we want to get these engineering teams considering. And again, it's something that they may have a blind spot, they may have an area where they hadn't even thought about. And that's what we really want to provoke them, is to think about the full ramifications. That's what it is to live up to our responsible principle as well as our robust principle, those two engineering principles, is education. And then they're excited, then they want to lean into fulfilling those principles.

>> Help us understand how we're going to... You mentioned engineers love solving problems. So help us solve this problem on power and cooling. What's the value chain or the value contribution of solutions? We're hearing silicon companies, compute companies, liquid cooling on that one side. On the other side, there's new forms of energy and physically locating in places where you can use ambient cooling, et cetera, et cetera. So connect those two endpoints for us.

Kirk Bresniker

>> And every time we take a joule of energy in, high-quality electricity, fantastic, we can do so much with that, we put out a joule of heat, low-quality energy. What happens in between? The question is how is that beneficial?

>> It sounds like algebra.

Kirk Bresniker

>> It is, it's energy in and the conservation of energy, but what did you do? What did you accomplish with that? And that's really what we're trying to understand and design for is accomplishing the most societal benefit. And so if you come over and we invite people over to talk to our team. The lab's demos, part of what that is is that exascale data center creating a living digital twin of the entire energy ecosystem around that. So you can not only make the best possible decisions, but do it transparently, so you can show your work. And so when we look at the novel technologies, we say, "Okay, great liquid cooling, that is the best solution through the physics, the economics." And then you can bring in the societal element and demonstrate how you're achieving value, how you're achieving benefit, how you've been the best possible steward of that energy. Because energy is the potential to do work. And that's what we need to understand. Are we doing the right work? The right technology, used the right way.

>> So liquid cooling is a topic that came up certainly at this keynote, and it came up a lot in the earnings call that Antonio was... after the earnings print. And Antonio's talking about direct liquid cooling as a differentiator for HPE and your decades of experience in supercomputers. I remember I was at GTC... I have a friend who runs a company called Omni Services, you've never heard of it, but they make hoses. And I was at GTC and I saw this supercomputer with all these hoses. I took a picture of it, I said, "Scott, you should be in this business." He said, "We're looking into this business, we're strategy planning." And he was explaining that things are going to change dramatically in terms of the infrastructure for liquid cooling. And direct liquid cooling is something that you guys have an advantage in. Can you explain how differentiable that advantage is and why it's so difficult?

Kirk Bresniker

>> And it comes down back to that physics equation, how do I move this heat? For a long time we thought no one's going to make a processor over 100 watts or 300 watts or now 500, 750, 1,000, 1,500 watts. All of that heat has to be removed, it has to go somewhere. And to date, the way you did it is you had a big old fan, you blow some air over it. Air is cheap, air is great, it's non-conductive, it does a pretty good job until you reach that threshold, until you reach that breaking point. And that's where the physical advantage of taking out that heat by transferring it into liquid. And then the great thing about liquid is then you can pipe it off somewhere. You can pipe it off to a greenhouse, you can pipe it off to a district heating solution. You can do a lot more because of that heat capacity of the fluid moving and controlling it. But it's actually challenging. It's challenging to have water and electricity in the same place at the same time. And that's where the decades of understanding the hundreds of patents that we've accumulated, that practical working knowledge, how do you solve the physics and economic equations of direct liquid cooling when it says, "That's what you need to do," that's the first step. Knowing how to do it, how to do it well, and how to have it work day in, day out for decades is the knowledge that we've accumulated.

Rebecca Knight

>> So Kirk, as we've said, it is Juneteenth. The nation is in mourning over the passing of Willie Mays, and we're talking a lot here about human rights and getting to the right outcomes. Later today at HPE Discover there's going to be a tech talk by a photojournalist, Lisa Kristine. She has traveled the world taking pictures of enslaved people in different countries. She's put together a beautiful stirring-

>> Can I show this?

Rebecca Knight

>> Yes, absolutely... a book. And we talk a lot about AI and putting AI to work and solve these immense challenges that we are facing. How do you think about AI's approach to solving something like human trafficking?

Kirk Bresniker

>> And so for us, it is perhaps the most potent tool we have to be working against human slavery. What we found in our own labor supply chain, the price of removing enslaved labor from your supply chain is eternal vigilance. You cannot close your eyes for a minute. So you have to continuously be examining... and that's what AI can give us, that longitudinal view over all of these signals to find out, to find those hidden hints in, maybe it's in the transactional record, that's where we are working with our NonStop team. Maybe it's in satellite reconnaissance looking for, "Why is there a CO2 plume in the middle of this desert?" Well, it might be because there's an illegal brick factory that's there, and that CO2 plume that you're seeing is the hint you need to then send in the reconnaissance team to try and understand. So a very potent tool for us to look for those hidden signs, those hidden signals. Of course, we also have to think about the fact that we're not the only ones. The adversaries are going to have these tools too, and they can use these tools to hide better. So it is a continuous challenge and one that we are committed to doing. And because it's, again, part of our legacy is establishing the industry's leading practices on labor supply chain, and certainly Lisa's work... And we had her with us at Davos, it was a fantastic discussion. And certainly John Schultz, our chief operating officer, this is his personal mission, is to really establish us as the leader across all of the industries in understanding how do we find the hidden signals of enslaved labor.

>> I know we're out of time, but real quick, you grew up in Silicon Valley, right?

Kirk Bresniker

>> Yes.

>> Was there a book that inspired you? I know I was on the East Coast, it was the Soul of a New Machine in the eighties. Was there a book, like a jobs book, or anything that inspired you to get into technology?

Kirk Bresniker

>> For me, my dad grew up in the East Bay, youngest of 11 on a chicken ranch in Hayward. In 1959 he graduated from high school, Hayward High. They were the Hayward High Farmers. The Agrarian was their yearbook. And it's hard to imagine that Silicon Valley used to be the valley of harsh life and an agricultural center, but he was the first one in his family to go to college, engineering degree. And what inspired me was when we were little, he went back and got his masters, early bird classes, goes to work, comes home, does his homework at night, and looking over his shoulder he had his engineering handbook with all of the tables of integrals and all the log tables. He had the slide rule my mother gave him-

>> Not even an HP calculator?

Kirk Bresniker

>> No, this was before the calculator. But my mother, she took her savings from her job at Capwell's Department Store down in Oakland, bought the most expensive slide rule she could afford as an investment in their future because it made him a better engineer. He used it every day until he got that HP calculator.

>> It's like, "Slide rule, look it up if you don't know."

Rebecca Knight

>> There you go.

Kirk Bresniker

>> But yeah, but looking over his shoulder, looking at the problems he was doing late at night, and then he'd go to school the next morning, that was what inspired me.

>> He was your idol, that's awesome.

Rebecca Knight

>> Kirk, always a pleasure having you on. That's a great story. That's great-

Kirk Bresniker

>> Thank you.

Rebecca Knight

>> Please stay tuned for more of theCUBE's live coverage on Juneteenth at HPE Discover. I'm Rebecca Knight for Dave Vellante. You are watching theCUBE, the leader in enterprise tech news and analysis.

Kirk Bresniker returns to theCUBE to dig deeper into the responsible design and use of AI

Rebecca Knight

>> Wow. I tell you, I needed the eyedrops. Willie Mays.

Rebecca Knight

>> Willie Mays, yes-

>> Say, "Hey, kid."...

Rebecca Knight

>> a legend.

>> Four decades the guy played baseball, amazing.

Rebecca Knight

>> Yes.

>> I never got to see him play, but my dad did.

Rebecca Knight

Kirk Bresniker

>> Thanks guys. Great to be here, thank you.

>> Good to see you.

Rebecca Knight

>> Yeah. Yeah, so responsible ethical AI, that is something that we're talking a lot about here at HPE Discover, but it's something that actually you've been talking about for many, many years.

Kirk Bresniker

Rebecca Knight

Kirk Bresniker

>> So you started with first principles and then what Kirk, then it became just modeling the permutations. Did it become a math problem, a social science problem?

Kirk Bresniker

>> It sounds like algebra.

Kirk Bresniker

Rebecca Knight

>> Can I show this?

Rebecca Knight

Kirk Bresniker

>> I know we're out of time, but real quick, you grew up in Silicon Valley, right?

Kirk Bresniker

>> Yes.

Kirk Bresniker

>> Not even an HP calculator?

Kirk Bresniker

>> It's like, "Slide rule, look it up if you don't know."

Rebecca Knight

>> There you go.

Kirk Bresniker

>> But yeah, but looking over his shoulder, looking at the problems he was doing late at night, and then he'd go to school the next morning, that was what inspired me.

>> He was your idol, that's awesome.

Rebecca Knight

>> Kirk, always a pleasure having you on. That's a great story. That's great-

Kirk Bresniker

>> Thank you.

Rebecca Knight