Data Protection & AI Summit | Data Protection & AI Summit AnalystANGLE w/ Scott Hebner

Clips
News
More from Data Protection & AI Summit

Christophe Bertrand

Principal Analyst

SiliconANGLE & theCUBE

Scott Hebner

Principal Analyst, AI

theCUBE Research

Data protection: The missing link in AI readiness

In the generative AI gold rush, it’s easy to forget that no amount of modeling magic can compensate for low-quality data. As enterprises race to deploy artificial intelligence, they often overlook the less glamorous but utterly essential work of securing, governing and preparing their data foundation. That omission is more than an operational oversight; it’s a strategic liability, especially when it comes to data protection for AI.“The lifeblood of AI is data,” said theCUBE Research’s Scott Hebner (pictured, right). “There’s no AI without an … information architecture. The data is the critical component of any AI system, and so that data has to be of the highest quality

Three insights you might have missed from theCUBE’s Data Protection & AI Summit

AI-powered data protection has become both a powerful tool and a growing necessity in today’s threat-filled digital landscape. IT leaders around the globe have come to recognize that AI and data security are strategic imperatives, making efficient storage and effective governance critical.Subjects such as AI-powered data protection were of key focus during theCUBE Research’s recent Data Protection & AI Summit. AI can be very efficient in attacking one’s environment, meaning organizations must fight AI with AI in many ways, according to Christophe Bertrand, principal analyst at theCUBE Research

Data protection: The missing link in AI readiness

In the generative AI gold rush, it’s easy to forget that no amount of modeling magic can compensate for low-quality data. As enterprises race to deploy artificial intelligence, they often overlook the less glamorous but utterly essential work of securing, governing and preparing their data foundation. That omission is more than an operational oversight; it’s a strategic liability, especially when it comes to data protection for AI.“The lifeblood of AI is data,” said theCUBE Research’s Scott Hebner (pictured, right). “There’s no AI without an … information architecture. The data is the critical component of any AI system, and so that data has to be of the highest quality

play_circle_outline Importance of data protection in AI projects for successful implementation and recovery.

play_circle_outline Unlocking AI Potential: The Impact of Quality Data on Performance and Overcoming Challenges in Data Utilization

play_circle_outline Emphasis on proactive data protection design to mitigate business risks in AI.

play_circle_outline Leveraging AI for Effective Data Management and Enhanced Cyber Threat Protection in Business Functions

Info
Transcript

Data Protection & AI Summit AnalystANGLE w/ Scott Hebner

Christophe Bertrand

Principal Analyst SiliconANGLE & theCUBE

HOST

Scott Hebner

Principal Analyst, AI theCUBE Research

HOST

In this Analyst Angle segment from the Data Protection + AI Summit, Christophe Bertrand, principal analyst at theCUBE Research, sits down with colleague Scott Hebner to map the new data protection playbook for AI driven enterprises. The conversation examines why there is no AI without a resilient information architecture and dissects the two sides of the same coin: protecting data for AI use and harnessing AI to fortify cyber resilience.

Hebner notes that 95 percent of enterprise data is still off limits to AI and only 38 percent of leaders trust the d... Read more

explore Keep Exploring

What key considerations should be taken into account when embarking on AI projects, especially regarding environmental protection and data management? add

What are some challenges organizations face in utilizing their data for AI projects? add

What should be considered when developing AI infrastructure and ensuring data protection? add

What are the potential roles of AI in managing and securing data for organizations? add

bolt Powered by CUBE AI

Data Protection & AI Summit AnalystANGLE w/ Scott Hebner

search

Christophe Bertrand

>> Hello everyone and welcome to the Data Protection & AI Summit. My name is Christophe Bertrand. I'm a principal analyst here at theCUBE Research, and this session is what we call an analyst angle. I'm going to talk about what's going on in the industry with one of my colleagues, Scott Hebner. Scott, welcome.

Scott Hebner

>> Thank you so much for having me.

Christophe Bertrand

>> So Scott is our principal analyst focused on ai, and of course, well, I focus on data protection, cyber resilience, data management. So we're going to chat about what's going on and the multiple facets of data protection and AI and what this is all about. So Scott, let me start by asking you a simple question, which is, why should people worry about data protection when they think about AI? What's your take?

Scott Hebner

>> Well, as you know, the lifeblood of AI is data. There's no AI without an IA, IA being an information architecture. So the data is the critical component of any AI system. And so that data has to be of the highest quality. It has to have integrity built into it on how it's processed, how it's used, and most importantly, it's got to be protected. Both regulatory and from various threats that may be out there. So any good long-term AI strategy is going to start with your data layer, and that includes data protection.

Christophe Bertrand

>> Absolutely, and this is very important because what you've just said I think is fundamental and in my opinion today, a lot of people embark on AI projects, infrastructure projects, data projects, lots of different initiatives are in play, but they may not be thinking about how to protect and how to recover the environment should it be necessary. So really in terms of what's going on in the industry, what we're going to be looking at throughout the summit is two sides of the same coin. So what we were just discussing, how do you protect the data essentially for AI purposes? But also how you can leverage AI in the context of data protection. So we have a number of great speakers throughout the summit who will cover both topics with us. Scott, let's talk about what you first identified for us, which is, okay, there is no good AI without essentially a good information architecture and really data that is protected. So what does that take from your standpoint? Where does it start? What do you see as maybe some of the shortcomings today that would make you say, well, we're just getting started, we're not ready yet?

Scott Hebner

>> Well, if you look at the vast majority of the data that an organization sits on, the proprietary enterprise data, very little of it is actually used today in general, and even less is used in their AI. And when you kind of dive a little bit deeper into these AI projects, you find out that it's because they're not sure how to protect it, right? It's a risk assessment. So the data's there, but it's not ready to be used for AI. And we estimate some 95% of an enterprise's data is just simply not ready. And protection's a big part of that. So what's happening is AI models, AI agents, things are getting created out there, but they're only using a subset of the data it really needs because it's simply not ready, and therefore it's not quite as effective of an AI capability as it could be. So that's why it's so critical to start there. And I've done a lot of work over the years and analyzing why some AI projects fail, and invariably it comes down to the data. Some organizations have built the most sophisticated, eloquent AI capabilities, but the data that feeds it wasn't good enough, didn't have the integrity, or there wasn't enough of it. So I think it's a huge challenge for any organization that wants a high integrity, a high productivity solution around AI.

Christophe Bertrand

>> And you've mentioned the fact that the data has to be protected. So if we break this down, there are multiple issues at play here. But look, we know that AI is happening. Projects are going to get better, things are going to get better, but what you need to think about is really protecting your data assets as you think about you want to make them reusable or intelligent for data reuse for AI. So you have multiple trends intersecting each other that I want to break down. So the baseline is this, AI is important to you as a workflow, it's important to you as a workload, and important to you as a business. For this reason, anything that is part of AI infrastructure has to be protected. I think that's the baseline. Don't make it a second thought. It is actually a design requirement. I think the reason why issues come up is because that's not something that's considered. And by the way, we've seen that in many new workloads or workflows throughout the years and waves of technology. The other piece is, and that's where the intersection starts, is also the data has to be actually okay to use or reuse, which is where we get into governance and compliance questions. By the way, compliance very, very often means two things. It means you have to have a backup to be able to recover, and it means you have to be cyber resilient too. So that's the third intersection, cyber resilience. So we get essentially a perfect storm here for a lot of organizations trying to really think about ai, and that's why we have this summit because we believe that data protection is a fundamental stone or foundational stone to being able to build an infrastructure that is AI friendly, AI compliant, and AI resilient as well. The other part that I want to talk about, and you just mentioned Scott, the agentic component. Okay, in time agents are going to be acting almost as individuals manipulating data, doing stuff with IT for business outcomes, for process outcomes, maybe talking to each other, making decisions for the business. What is the risk here for data itself? Could agents go nuts or be misconfigured and essentially destroy data?

Scott Hebner

>> Well, yeah, I think you're upping the stakes when you move into agentic AI. Most businesses today would have either predictive models that are going to make a prediction of what may happen in the future based on their past data, what has happened in the past, or they're using generative AI assistant that is basically designed to automate tasks, right? Repeatable tasks, to create content, to analyze information. Those are relatively low risk activities for people to do within a business. With agents though they're supposed to help you pursue a goal, help you make actual business critical decisions, help you solve a problem and even act autonomously. So when you get into the world of decision-making, the stakes become much higher. And you also got to keep in mind that these AI systems are just going to create more data that has to then go back into the life cycle. So data protection I think is not just you scan the data, you make sure it's what you want it to be, and then you go away. It's got to be continuous. It's always being fed here. So I think, again, I almost think of it as the bloodline. It flows through the AI system and you got to keep your blood healthy and healthy means in large part protected. So it's a living organism, if you will. It always has to be monitored, and that's where I think a new phase of governance is be implemented here. And I'm sure we'll get into the volumes of data which make it even harder soon, right, Christophe?

Christophe Bertrand

>> Absolutely. So actually it's interesting because we were chatting on your podcast recently and we'll provide links to our viewers, this great podcast. We were talking about governance, we're talking about also the volume of data that will be created. So again, I think people have to understand here, everything's fun and games until you add at scale at the end of the sentence, and this is a scale game. Scott, if I'm not mistaken. We are looking, in the next few years, at hundreds of zettabytes of new data, very likely being generated by AI itself. So as we think about protecting what we already have to be able to support current initiatives, and clearly there are lots of issues already. There is more to come and there is more to go protect. What were those numbers at a high level? Actually didn't me bring up maybe a quick chart here that we have that I think you sourced from a number of sources, SAP, McKinsey, IDC, number of organizations. What does it tell us exactly?

Scott Hebner

>> Well, this is telling us, and this is a projection based on those sources. That if you look at the newly generated data within an organization, so this does not include what already exists, you can see just a rapid growth each year. You're starting to approach exponential here. And at the same time, if you think about it, only so much human involvement and traditional governance platforms can handle that kind of growth. So that's going to lead us to, well, you need to apply AI to help actually govern the data. But more importantly on this chart I think is the red line that today only about 38% of business leaders and users trust the data that they're making decisions off of or doing analytics on. So that's just whether it's perception or reality, in some ways it doesn't matter, only 40% of them actually trust the data. If you kind of build that out over time, and if the data newly generated data grows at the rate of the blue bars, you can extrapolate here that the trust in that data is just going to nosedive here. So you could potentially get to 2030, and have less than 20% of your stakeholders and users trusting the data they're using unless something's done differently.

Christophe Bertrand

>> And this is a topic that definitely we'll pick up on in the next few weeks in another summit around data governance and AI. But the point being here, just looking at protecting 277 zettabytes, and who knows if that number is even going to be correct, is probably going to be more, if you want my opinion. It becomes something that you must think about now. Because the truth is, data that is not protected is data that is very likely not compliant and you introduce business risk. So this is the part where, okay, I have to protect this AI infrastructure. And again, we'll have some great discussions. I will just hint that one of the conversations we will have is around tape for example, because how do you think you're going to store all of that data and you'd better have a great medium to do so. And we have also great sponsors and vendors who are coming today to talk about other aspects associated with AI. So I'd like to switch to that with you, Scott, and get your perspective because I mentioned those three big trends, intersecting data management, compliance, governance, protection for the purpose of recovery and also cyber resilience. And clearly AI can actually be your friend if you think about all of the investments being made now on the vendor side to automate, to make solutions more efficient, more operationally successful at protecting data, there is a positive aspect to AI and I expect agent-IC AI will be part of that as well. So what's your take on how AI can help better fight the cyber attackers who themselves probably use AI, help vendors who provide recovery capabilities and services better provide those services at scale, given the zettabytes we're looking at, what's your take on this and what do you think the architecture may look like in terms of what these organizations decide to go for?

Scott Hebner

>> Well, Christophe, as you and I have been talking a lot. AI can become a friend or a foe, and it's probably going to be both. So you got to keep an eye on the foe part, which is more sophisticated threats, and sometimes you got to fight fire with fire. So you're going to need AI to protect against those threats. But at the same time, I think if you just think about the volumes of data getting generated that we talked about, you're going to need AI to be able to efficiently process all that data, sort through it and determine how to protect it, how to ensure the integrity of it. And if that data is always changing, the life cycle of the value of data is shortening every day that goes by. And so it becomes a very dynamic environment. And I think first of all, AI is the way to deal with these very, very large volumes of data and the complexity of how that data is structured. And I think the second thing is it's going to learn from that data and start to understand what is really quality data, what is protected and what is not. And then it will learn from that and come back and do it better next time. So as the system builds over time, it becomes better and better at what it does. And so I do think building in a ai, let's call architecture that underpins governance and trust management, those frameworks including protection and regulatory compliance and corporate policy is really, sometimes you wonder if it becomes table stakes. I mean, it's just something that you have to do otherwise how are you going to keep up with all this? It's moving at warp speed here. So there's got to be like we showed before on that chart you showed there's got to be a new way of doing it, otherwise you're simply going to fall behind and maybe never be able to catch up.

Christophe Bertrand

>> Exactly. And I think that's risk. And more importantly it's business risk because that means you have non-compliant systems, non- compliant data potentially. You can also be attacked again from the outside with prompt injections trying to get into your own AI processes. So what we're seeing here is this snowball effect. There's AI creating more AI data, data that has to be protected from by design, but also a great possibility or set of great possibilities with AI becoming your friend to be able to manage this amount, enormous amount of data that is being created and will be created in the next few years. I mean, we're talking projections that are only 10 years, not even 10 years out, and it's multiplying by significant amounts. So one thing that I know that we've talked about in the past, and I'd like to share today as we just set the stage for just the broader picture, the bigger picture is the chain, what we call the chain link of challenges. And that's definitely a model that you've proposed here, four stages that organizations can actually use data identification, governing at scale for access, and of course with the objective that in the end what you get is trusted AI so that you can mitigate your risks. So let's talk about these stages briefly, Scott, from your standpoint. I would say data protection, we should maybe add it on the left-hand side, but it's really throughout the four stages you have to have well-protected data processes, agents, and of course all the way throughout not only protected but compliant, bearing in mind that most of the data, as you point out, is not ready and most of it is not even compliant for AI. So walk us through this from your standpoint.

Scott Hebner

>> On the left-hand side, I think it starts with what data is most important, because perhaps not all of it is. And for any high dimensional space, like if you're doing AI for financial services or life sciences or even for very complex international operations, different aspects of data have different influences on the outcome. So I think identifying the data, classifying it, making sense of it if you will, and really sorting through what is most important. It's not like the human brain when you're in REM sleep and you're sleeping, it sorts through all the information you took in during the day and it kind of stores only what it considers to be the important things. So I think it starts with that. The scale governance is then just because you've identified it and classified, it doesn't necessarily mean that it's ready to be used. And I think Christophe, that's where the protected data comes in because it's only governed if it's protected. And you're scaling the ability for people to use it. Whether it be an AI model, an AI agent, a human being, a business analyst, a business partner, whatever it might be, you have to be able to scale all that. And when I kind of look at AI use cases so far, I mean most of it's been dominated by marketing and customer service, but when you look in the area of IT, the use of AI and application development has really scaled up over the last couple of years followed by IT operations. I think data protection is probably the next area of scale up within IT of using AI. And that's going to do a lot. I think too, as we've just been discussing, the two chain links or circles on the left-hand side of the chart here. And then once you have the bloodline healthy and you get your metabolic tests at Quest or whatever and everything is, your data is great, there's still challenges on the other side of the coin, which is now I got to have trusted AI models. And that's a whole bunch of different things. That could be, how does user trust it? Is it explainable? Can you interpret it? Can you trace it? Can you audit how it's making decisions or whatever? And then of course, everyone is now operating in this incredibly fast-moving regulatory landscape. I think we talked last time, Christophe, that seventy- plus countries now have over a thousand different regulations and they're growing by 20, 25% every year and they're always adapting. So how do you keep track of all this stuff? I think that's the concept that we came up with together here about this chain link. Solving one of them's not going to do it. You really got to think about all four of these at a minimum. Otherwise, you're going to break the chain of value or protection.

Christophe Bertrand

>> Exactly. And that's why, I mean really if there is one statement everybody has to remember, it's really two sides of the same coin, which is you need strong data protection principles in place in order to really benefit and launch into AI projects. It's really the baseline has to be thought through, has to be done by design. You have to protect against the cyber attackers who are themselves using AI. And that's why you have to find solutions that are really appropriate for that. That may be themselves, and we'll talk about that, use AI to really help detonate all these threats to analyze them to better protect you. Of course, you have to be able to recover all of the components. And by definition that helps in making you compliant. Now we'll see what the regulations really end up doing, and that's a topic we'll cover. I think it's at the end of September in our data Governance and AI Summit. So stay tuned for that in the next few weeks. But I think this will be a very interesting set of conversations because everything is intertwined. I think those three big trends that I mentioned, cyber resiliency, data management, and by the way, unstructured data, unstructured data, plenty of it, plenty that needs to be governed, plenty that you're going to need for AI. And of course data protection. All of these components are merging, combining. So there's definitely lots here to cover. And that's to be able to fully get to that level of AI you can trust. AI that will very likely change your business, but there's a lot of work ahead and that's what we want to do. The Summit is one of the first steps in helping you getting to this AI Nirvana. It's going to take a while, but let's get data protection done right. Let's understand how to leverage AI and data protection, how to protect AI itself. And if we can do that and expose some of the great solutions that exist today, I think it will be really good. It will be definitely a good day and time well spent. Scott, I'd like to thank you so much for your insights. They're invaluable. And we'll talk again in the next few weeks in another Summit about the governance components.

Scott Hebner

>> Thanks so much. It's always great talking with you.

Christophe Bertrand

>> Thank you everybody, and we'll see you on the next session. Stay tuned. There's plenty to cover here in the Data Protection & AI Summit. My name is Christophe Bertrand, principal Analyst at theCUBE Research. Thank you.