3042: AI Governance and Risk Management Insights from Securiti AI

[00:00:04] [SPEAKER_00]: How can C-suite executives effectively balance the push for innovation with the critical

[00:00:10] [SPEAKER_00]: need for security and compliance?

[00:00:13] [SPEAKER_00]: Especially when you throw in the fact we're in an AI driven world right now.

[00:00:18] [SPEAKER_00]: Well today Jack Berkowitz, Chief Data Officer at Securiti AI is going to explore this complex

[00:00:25] [SPEAKER_00]: question and challenge with us today.

[00:00:28] [SPEAKER_00]: Jack brings with him over 30 years of experience guiding Fortune 500 companies through the

[00:00:33] [SPEAKER_00]: intricacies of AI governance and data security, and with data transforming industries and

[00:00:39] [SPEAKER_00]: regulations tightening, how can organisations stay ahead while ensuring their data is secure

[00:00:46] [SPEAKER_00]: and compliant?

[00:00:48] [SPEAKER_00]: Well today we're going to dive into the proven strategies for AI risk management, the role

[00:00:54] [SPEAKER_00]: of Securiti AI's Data plus AI Command Centre, and how transparency can build trust in AI

[00:01:03] [SPEAKER_00]: systems.

[00:01:04] [SPEAKER_00]: Are you or your business thinking of launching a podcast in 2025?

[00:01:09] [SPEAKER_00]: Well dreams are about to become a reality because I have three amazing options that will guarantee

[00:01:15] [SPEAKER_00]: that you can make that happen.

[00:01:17] [SPEAKER_00]: Option one, you record and I handle the rest.

[00:01:20] [SPEAKER_00]: All you need is 30 minutes to go and record an episode, I'll take care of editing, setting

[00:01:26] [SPEAKER_00]: up the RSS feed and publishing it across all major podcasting platforms.

[00:01:30] [SPEAKER_00]: For just $2,000 you can work alongside me, I'll have you up and running with a season

[00:01:36] [SPEAKER_00]: of 12 polished episodes ready to go live.

[00:01:39] [SPEAKER_00]: So if you or your business are interested in getting a podcast live and you want to

[00:01:43] [SPEAKER_00]: work directly with me, visit my website techblogwriter.co.uk, email me techblogwriter at outlook.com, send

[00:01:51] [SPEAKER_00]: me a DM on any social platform such as LinkedIn, X or Instagram, I'm just at Neil C Hughes

[00:01:57] [SPEAKER_00]: on every single platform.

[00:01:59] [SPEAKER_00]: Option two, if you're looking for extended end to end support, well I've partnered with

[00:02:04] [SPEAKER_00]: AIpodcast.ing, they provide everything from website management to promotional material

[00:02:12] [SPEAKER_00]: and help improving engagement and you can also get 25% off your first month with the

[00:02:18] [SPEAKER_00]: code NeilCHughes and let them handle all the heavy lifting for you.

[00:02:23] [SPEAKER_00]: But option three, if you want to do it all yourself, LipSyn, the podcast host that I

[00:02:28] [SPEAKER_00]: use for this show, they are offering up to two months of free hosting when you use the

[00:02:32] [SPEAKER_00]: promo code TBW.

[00:02:34] [SPEAKER_00]: For example, if you were to sign up on November the 1st, you can start 2025 strong and without

[00:02:40] [SPEAKER_00]: paying anything till January the 1st, 2025.

[00:02:44] [SPEAKER_00]: Whichever path you choose, I'm here to make sure that you start the new year with a podcast

[00:02:49] [SPEAKER_00]: ready to take the world by storm.

[00:02:51] [SPEAKER_00]: Let's do it together.

[00:02:52] [SPEAKER_00]: But enough from me, let's get the guest on now.

[00:02:55] [SPEAKER_00]: So a massive warm welcome to the show, Jack.

[00:02:58] [SPEAKER_00]: Can you tell everyone listening a little about who you are and what you do?

[00:03:03] [SPEAKER_01]: Sure.

[00:03:03] [SPEAKER_01]: My name is Jack Berkowitz.

[00:03:04] [SPEAKER_01]: I'm the chief data officer of a startup company, later stage startup company called Security.AI.

[00:03:11] [SPEAKER_01]: And we work with businesses that are trying to deploy data systems and AI systems in a

[00:03:19] [SPEAKER_01]: secure way.

[00:03:20] [SPEAKER_01]: My entire job is actually a little bit different.

[00:03:24] [SPEAKER_01]: I'm working with the C levels of Fortune 500 companies having been at the C level previously

[00:03:31] [SPEAKER_01]: in a couple of big companies.

[00:03:34] [SPEAKER_00]: Well, it's a pleasure to have you on the podcast.

[00:03:36] [SPEAKER_00]: I'm looking forward to find out more information about the kind of things that C-suite are

[00:03:40] [SPEAKER_00]: talking about because there's so much hype around AI this week alone that we're recording

[00:03:44] [SPEAKER_00]: this podcast.

[00:03:45] [SPEAKER_00]: It feels like, yes, there's guys like me and you that talk about it all day long, but it

[00:03:49] [SPEAKER_00]: feels like it's entered the mainstream this week with the big announcements with Apple

[00:03:53] [SPEAKER_00]: and Apple Intelligence.

[00:03:54] [SPEAKER_00]: There was the Oprah interview last night, which seems to be put in front of everybody right

[00:04:00] [SPEAKER_00]: now.

[00:04:00] [SPEAKER_00]: But with your extensive experience, how do you see the role of the C-suite evolving and

[00:04:06] [SPEAKER_00]: balancing and the opportunities of AI innovation with the growing complexities of security

[00:04:12] [SPEAKER_00]: and compliance?

[00:04:13] [SPEAKER_00]: It's a real mixed bag and there's so much going on, but what are you seeing here?

[00:04:17] [SPEAKER_01]: Well, it's really that balance, right?

[00:04:20] [SPEAKER_01]: When you're sitting there and you're talking about strategy, not strategy over six months,

[00:04:24] [SPEAKER_01]: but strategy over three to five years, every C-suite's looking at the business advantage

[00:04:30] [SPEAKER_01]: or the business disruption that these technologies are going to have.

[00:04:33] [SPEAKER_01]: If you look at just the reasoning stuff that OpenAI released yesterday, that disrupts huge

[00:04:42] [SPEAKER_01]: industries around SaaS applications or any sort of services type work or knowledge-based

[00:04:50] [SPEAKER_01]: work.

[00:04:50] [SPEAKER_01]: At the same time, everybody's worried about, well, wait a second, this is about intellectual

[00:04:56] [SPEAKER_01]: property, but it also may be about information about our employees, information about our

[00:05:01] [SPEAKER_01]: relationships with our customers, information about our earnings release and how do we make

[00:05:08] [SPEAKER_01]: sure that we can operate as a company or as an organization, respecting both the regulations,

[00:05:16] [SPEAKER_01]: but also just respecting the people we work with.

[00:05:20] [SPEAKER_01]: While at the same time, getting ourselves in a position to react to these pretty fundamental

[00:05:27] [SPEAKER_00]: changes.

[00:05:28] [SPEAKER_00]: Yeah, there's so much happening around this at the moment.

[00:05:30] [SPEAKER_00]: I think it was in February of this year, I spoke with your CEO, Rehan.

[00:05:35] [SPEAKER_00]: I think we were talking about securing enterprise data in this generative AI landscape that we

[00:05:42] [SPEAKER_00]: find ourselves.

[00:05:43] [SPEAKER_00]: But at SecureAI, you are leading efforts in data security and governance.

[00:05:48] [SPEAKER_00]: I remember from our last conversation, but can you just explain how data and AI command

[00:05:53] [SPEAKER_00]: center that you've got there, how that is helping organizations securely manage their AI and

[00:06:00] [SPEAKER_00]: their data?

[00:06:01] [SPEAKER_00]: And the reason I ask that, I think there's so much hype at the moment.

[00:06:03] [SPEAKER_00]: A lot of businesses looking about what they're going to do next, how they're going to do

[00:06:06] [SPEAKER_00]: this.

[00:06:06] [SPEAKER_00]: They need to do something, but knowing how to do it effectively, that's where things

[00:06:10] [SPEAKER_00]: often get a little bit complex.

[00:06:13] [SPEAKER_00]: So can you expand on that data plus AI command center that you've got there?

[00:06:17] [SPEAKER_01]: Yeah.

[00:06:18] [SPEAKER_01]: So I was chief data officer of a Fortune 500 company prior to joining it.

[00:06:24] [SPEAKER_01]: Our biggest issue wasn't as much knowing how to build an AI system.

[00:06:28] [SPEAKER_01]: I had a bunch of data scientists, we could go do that.

[00:06:31] [SPEAKER_01]: But it was really understanding all the data assets we had, where they happened to be,

[00:06:36] [SPEAKER_01]: how they were flowing through our systems.

[00:06:40] [SPEAKER_01]: And once we had that knowledge, well, wait a second, what about all this unstructured

[00:06:45] [SPEAKER_01]: data that the generative AI world is all about?

[00:06:48] [SPEAKER_01]: Where is it?

[00:06:49] [SPEAKER_01]: Have we been managing it?

[00:06:51] [SPEAKER_01]: Do we even know who's updated it and is it current?

[00:06:56] [SPEAKER_01]: And so the data command center idea is to provide a view of this information and then

[00:07:02] [SPEAKER_01]: put controls in place that can alert you.

[00:07:04] [SPEAKER_01]: So it can alert you if data has moved from the UK to the US inadvertently.

[00:07:10] [SPEAKER_01]: It can alert you if you violated a regulation.

[00:07:18] [SPEAKER_01]: It can alert you or it can actually help you take action to make sure...

[00:07:23] [SPEAKER_01]: One of the big things right now is around data access governance around SharePoint or

[00:07:28] [SPEAKER_01]: Office 365.

[00:07:30] [SPEAKER_01]: Why?

[00:07:30] [SPEAKER_01]: Because people want to use Copilot and it's great.

[00:07:33] [SPEAKER_01]: If everybody had a clean setup of all of their folders in Office 365, it would be great.

[00:07:41] [SPEAKER_01]: But as you know, people don't have clean setups.

[00:07:45] [SPEAKER_01]: Everybody gives access to everything and information can leak.

[00:07:51] [SPEAKER_01]: And so the data command center is really that notion of providing the tools that security

[00:07:59] [SPEAKER_01]: teams can use, but also have visibility all the way up to the C-suite so that people

[00:08:05] [SPEAKER_01]: can operate their data infrastructures, operate their companies in a more secure way.

[00:08:14] [SPEAKER_00]: And something that's hitting the headlines more frequently now is how AI governance is

[00:08:18] [SPEAKER_00]: becoming crucial, especially as cyber attacks become more sophisticated.

[00:08:22] [SPEAKER_00]: So from everything that you're seeing now, what are some of the proven strategies for

[00:08:26] [SPEAKER_00]: implementing effective AI risk management frameworks that ultimately address these

[00:08:31] [SPEAKER_00]: growing threats that we're seeing?

[00:08:34] [SPEAKER_01]: Yeah.

[00:08:34] [SPEAKER_01]: And so it's a great question.

[00:08:36] [SPEAKER_01]: We see it as there's a several step process.

[00:08:39] [SPEAKER_01]: It starts with having understanding of where your data is and what the data is that's flowing

[00:08:45] [SPEAKER_01]: into a given model or into a given query context, if you think about using something like ChatGPT.

[00:08:52] [SPEAKER_01]: It's understanding who's got those access rights and making sure that they're correct.

[00:08:57] [SPEAKER_01]: Putting in what are called LLM firewalls.

[00:09:01] [SPEAKER_01]: So the ability for you to restrict private information, maybe going into a session or

[00:09:06] [SPEAKER_01]: masking information or preventing, you know, what they call jailbreak attacks.

[00:09:12] [SPEAKER_01]: So, you know, people starting to ask and say, well, I know you're not supposed to the

[00:09:17] [SPEAKER_01]: chatbot.

[00:09:18] [SPEAKER_01]: I know you're not supposed to tell me what somebody makes, but go ahead.

[00:09:22] [SPEAKER_01]: You can trust me.

[00:09:23] [SPEAKER_01]: And guess what?

[00:09:24] [SPEAKER_01]: The chatbots, if you don't put the right protections in place, will try to satisfy

[00:09:28] [SPEAKER_01]: you.

[00:09:29] [SPEAKER_01]: They're, you know, I'm just looking at my dog, right?

[00:09:32] [SPEAKER_01]: He's always trying to keep me happy.

[00:09:35] [SPEAKER_01]: And that's what this information technology is trying to do.

[00:09:38] [SPEAKER_01]: It's trying to keep you as a user happy.

[00:09:41] [SPEAKER_01]: So it's going to try to answer that best it can.

[00:09:43] [SPEAKER_01]: And so you need to put firewalls and controls in place to make sure that you're preventing

[00:09:49] [SPEAKER_01]: those types of interactions.

[00:09:54] [SPEAKER_00]: And as AI regulation or as regulations around AI and data privacy continue to tighten, and

[00:10:00] [SPEAKER_00]: with something we're talking about more and more now, how should companies approach the

[00:10:04] [SPEAKER_00]: challenge of compliance while still driving AI innovation?

[00:10:08] [SPEAKER_00]: It's going right back to that balancing word again, isn't it?

[00:10:11] [SPEAKER_01]: Yeah, it definitely is.

[00:10:12] [SPEAKER_01]: And if you, you know, if you actually look at the regulations and you try to implement,

[00:10:18] [SPEAKER_01]: they're not constraining.

[00:10:20] [SPEAKER_01]: They're actually the right things to be doing to build a system that you should be comfortable

[00:10:26] [SPEAKER_01]: with.

[00:10:26] [SPEAKER_01]: It's really about transparency, visibility.

[00:10:29] [SPEAKER_01]: It's really about making sure that you have an understanding of your information flowing.

[00:10:35] [SPEAKER_01]: It's making sure that you're checking your models are actually not, you know, inappropriately

[00:10:40] [SPEAKER_01]: introducing bias or things like that.

[00:10:43] [SPEAKER_01]: So the regulations, if you do it correctly, if you lay down the right foundations, you're

[00:10:48] [SPEAKER_01]: actually creating a springboard for your business advantage by complying with the regulations.

[00:10:54] [SPEAKER_01]: It's not that you, you know, yeah, you're going to avoid fines, but actually you're

[00:10:59] [SPEAKER_01]: going to put yourself in a better situation to, you know, getting back to the original

[00:11:04] [SPEAKER_01]: thing what the C-suite wants to do, right?

[00:11:06] [SPEAKER_01]: Which is take advantage of this technology.

[00:11:10] [SPEAKER_01]: So my thing would be is, you know, go ahead and look at a system, whether it's from security,

[00:11:17] [SPEAKER_01]: where I work or some of the other companies and start to build out those capabilities

[00:11:24] [SPEAKER_01]: and also the people that you're going to need to operate and maintain these systems and

[00:11:29] [SPEAKER_01]: go ahead and do it in an aggressive way because it actually will get you in a good position

[00:11:35] [SPEAKER_01]: for what you're going to be in the future.

[00:11:38] [SPEAKER_00]: And after doing a little research on you, I know you've spoken with numerous clients

[00:11:42] [SPEAKER_00]: about advancing AI and data implementations, but are there any real world examples of how

[00:11:49] [SPEAKER_00]: companies are enhancing their security through strategic AI governance that you're able to

[00:11:54] [SPEAKER_00]: share?

[00:11:54] [SPEAKER_00]: I appreciate you probably can't mention any names, but just to bring to life what we're

[00:11:58] [SPEAKER_00]: talking about, are there any examples you can share?

[00:12:00] [SPEAKER_01]: Yeah, I mean, I can't share specifics on a customer.

[00:12:04] [SPEAKER_01]: Yeah.

[00:12:05] [SPEAKER_01]: But, you know, I've been working on this for most of my career.

[00:12:08] [SPEAKER_01]: You know, I was looking in the mirror this morning and the beard is actually gray.

[00:12:15] [SPEAKER_01]: And you see these examples where, you know, by putting in the right security measures,

[00:12:21] [SPEAKER_01]: you're able to build customer support agents that, you know, drastically reduce the time

[00:12:29] [SPEAKER_01]: it takes to support a caller or move it from, you know, pure voice interaction onto, you

[00:12:37] [SPEAKER_01]: know, a chatbot.

[00:12:39] [SPEAKER_01]: And you see then increases in customer satisfaction.

[00:12:44] [SPEAKER_01]: You see reductions in churn.

[00:12:46] [SPEAKER_01]: And why is that?

[00:12:47] [SPEAKER_01]: Well, people put in the right security.

[00:12:49] [SPEAKER_01]: So the companies are competent to provide that capability and the customer, because,

[00:12:55] [SPEAKER_01]: you know, maybe falsely, but normally through building up trust over time, trust that system

[00:13:04] [SPEAKER_01]: to be used.

[00:13:05] [SPEAKER_01]: And it's really about trust, right?

[00:13:07] [SPEAKER_01]: If the customer starts asking things and all of a sudden their national identifier, their

[00:13:14] [SPEAKER_01]: – I used – when I lived in the UK, I had my national insurance number, right?

[00:13:20] [SPEAKER_01]: Or here in the US, your social security number.

[00:13:22] [SPEAKER_01]: If those numbers start showing up, then consumers or customers aren't going to trust, right?

[00:13:28] [SPEAKER_01]: But by masking that stuff, making sure that it doesn't leak out, making sure it doesn't

[00:13:33] [SPEAKER_01]: show up in the chatbot itself, people will build that trust and will start to use it.

[00:13:37] [SPEAKER_01]: And you start to see that advantage.

[00:13:39] [SPEAKER_01]: And that saves companies, you know, millions of dollars in costs, but also, you know, think

[00:13:45] [SPEAKER_01]: about it, by retaining your customers and keeping them happy, you know, massive acceleration

[00:13:51] [SPEAKER_01]: to your business.

[00:13:53] [SPEAKER_00]: And cyber security threats like data poisoning and model theft are also on the rise, is it?

[00:13:59] [SPEAKER_00]: Any advice you might have for any CISOs that could be listening who are re-evaluating their

[00:14:04] [SPEAKER_00]: security frameworks right now and trying to counter these evolving threats?

[00:14:08] [SPEAKER_00]: Any advice for those people listening?

[00:14:10] [SPEAKER_01]: Yeah, definitely.

[00:14:12] [SPEAKER_01]: So these systems don't work in isolation.

[00:14:14] [SPEAKER_01]: You can't just say, well, I've got a couple of smart data scientists, so you're going

[00:14:19] [SPEAKER_01]: to go work on that in the corner.

[00:14:21] [SPEAKER_01]: You know, first of all, you know, we're all familiar in the security world about data

[00:14:27] [SPEAKER_01]: poisoning attacks, but also, you know, just software supply chain attacks.

[00:14:33] [SPEAKER_01]: These things can happen in these models, right?

[00:14:36] [SPEAKER_01]: And so you need to have a system that's looking at it end to end from the data origination,

[00:14:42] [SPEAKER_01]: whether you're training the model or just using it for inference.

[00:14:44] [SPEAKER_01]: And most people aren't going to be training models.

[00:14:46] [SPEAKER_01]: Most people are just going to be using existing models for inference.

[00:14:50] [SPEAKER_01]: But you need to understand those.

[00:14:52] [SPEAKER_01]: You need to understand your data flowing in.

[00:14:54] [SPEAKER_01]: You need to understand how the firewalls are set up.

[00:14:57] [SPEAKER_01]: You need to understand it, whether or not your team's building it or you're using it

[00:15:02] [SPEAKER_01]: in a SaaS application.

[00:15:05] [SPEAKER_01]: You know, something like you're using in Salesforce or you're using in Workday or whatever it happens

[00:15:09] [SPEAKER_01]: to be.

[00:15:10] [SPEAKER_01]: And so I think, you know, an awful lot of CISOs that I've spoken to at the C level,

[00:15:19] [SPEAKER_01]: they understand this.

[00:15:20] [SPEAKER_01]: But when you go a level or two down in the organization, people are, because they don't

[00:15:24] [SPEAKER_01]: know the technology, say, oh, well, we have a data science team doing that.

[00:15:28] [SPEAKER_01]: And they'll worry about it.

[00:15:29] [SPEAKER_01]: And I think you really need to motivate your team and say, no, this is an us problem.

[00:15:36] [SPEAKER_01]: This is a we problem.

[00:15:37] [SPEAKER_01]: And all of us need to be working on this together as a company or as an organization to make

[00:15:44] [SPEAKER_01]: sure that we're operating in the most secure ways.

[00:15:47] [SPEAKER_00]: And from the outside, looking at your career, you've helped deploy AI and ML systems across

[00:15:52] [SPEAKER_00]: so many Fortune 500 companies.

[00:15:55] [SPEAKER_00]: I've got to ask, how do you ensure that AI systems are both secure and effective at scale,

[00:16:01] [SPEAKER_00]: particularly in industries facing high and regulatory scrutiny?

[00:16:05] [SPEAKER_00]: Because I hear a lot about companies that are struggling with this at the moment.

[00:16:08] [SPEAKER_00]: And if you can share around this?

[00:16:11] [SPEAKER_01]: Yeah.

[00:16:11] [SPEAKER_01]: So I always start, and the reason why I'm so excited to be at security is I always start

[00:16:18] [SPEAKER_01]: with this idea that we need to have observability and measurement.

[00:16:22] [SPEAKER_01]: You know, even, I was just relaying yesterday, so I spent part of my career at Oracle getting,

[00:16:28] [SPEAKER_01]: helping to get that, their first AI and ML projects off the ground with customers in

[00:16:35] [SPEAKER_01]: their business applications, like in ERP and HCM.

[00:16:38] [SPEAKER_01]: The first day we went out, everything was measured.

[00:16:43] [SPEAKER_01]: Everything about data access, model drift, everything was being measured, the first customer

[00:16:50] [SPEAKER_01]: to touch it.

[00:16:51] [SPEAKER_01]: It wasn't an afterthought.

[00:16:53] [SPEAKER_01]: And so not only were we confident in the business results that we were providing to

[00:16:59] [SPEAKER_01]: our customers, but we were confident in the security and the data flows and the visibility

[00:17:05] [SPEAKER_01]: throughout that system.

[00:17:06] [SPEAKER_01]: It took us years to build that out.

[00:17:09] [SPEAKER_01]: Today, you can take advantage of all those learnings from companies like ours, where

[00:17:17] [SPEAKER_01]: all of that's available, and you can just implement it and get going very quickly.

[00:17:21] [SPEAKER_01]: But to me, it's really about understanding and measuring proactively and not reactively.

[00:17:29] [SPEAKER_01]: It's being proactive and taking a stance and moving forward aggressively as opposed to

[00:17:36] [SPEAKER_01]: just letting it build up.

[00:17:38] [SPEAKER_01]: It's engineering debt or technical debt.

[00:17:41] [SPEAKER_01]: Letting that build up will just never get you to the other side.

[00:17:46] [SPEAKER_01]: So really, I spend an awful lot of time asking this one question, whether I'm at a board

[00:17:52] [SPEAKER_01]: of directors of a Fortune 500, whether I'm with the C-suite or whether I'm down inside

[00:17:58] [SPEAKER_01]: working with day-to-day hands-on keyboard people.

[00:18:04] [SPEAKER_01]: What's the definition of good?

[00:18:06] [SPEAKER_01]: Let's just get down to the one thing.

[00:18:09] [SPEAKER_01]: What would you consider to be good?

[00:18:12] [SPEAKER_01]: Good in terms of your business, good in terms of security.

[00:18:14] [SPEAKER_01]: If we can get agreement on that, then everything else just falls in line.

[00:18:19] [SPEAKER_01]: And I think that sort of stepping back, getting to the whiteboard and having that discussion

[00:18:25] [SPEAKER_01]: is the most important way to have both a successful program and a secure program.

[00:18:31] [SPEAKER_00]: And if we dare to look ahead, any steps that organizations should be taking to operationalize

[00:18:37] [SPEAKER_00]: AI transparency, again, another word we're hearing a lot about, but also trust so we

[00:18:43] [SPEAKER_00]: can improve adoption and achieve those long-term business goals, especially in rapidly changing

[00:18:49] [SPEAKER_00]: digital landscape that we're seeing so many big changes right now.

[00:18:52] [SPEAKER_00]: And everyone's rushing forward with AI.

[00:18:54] [SPEAKER_00]: It is a big, exciting technology, but it's all about business value and achieving business

[00:18:59] [SPEAKER_00]: goals.

[00:18:59] [SPEAKER_00]: That's what we need to get back to.

[00:19:01] [SPEAKER_00]: So any advice here?

[00:19:04] [SPEAKER_00]: Yeah.

[00:19:04] [SPEAKER_01]: I think one of the things that's very practical item is write down your principles.

[00:19:12] [SPEAKER_01]: Write down how you intend to operate, how you intend to introduce AI and put it out

[00:19:21] [SPEAKER_01]: on your website or put it out in your materials so everybody can see it, whether it's your

[00:19:27] [SPEAKER_01]: employees, whether it's your customers, whether if you're regulated, it's your regulators.

[00:19:32] [SPEAKER_01]: Write these things down.

[00:19:33] [SPEAKER_01]: I was at ADP and in 2019, we wrote out our data and AI governance and security principles

[00:19:44] [SPEAKER_01]: and published that.

[00:19:45] [SPEAKER_01]: There were only five principles.

[00:19:47] [SPEAKER_01]: I think now it's seven principles.

[00:19:50] [SPEAKER_01]: We wrote it out and we said, hey, this is how we're going to operate as a company.

[00:19:55] [SPEAKER_01]: That transparency then drives a lot of decisions.

[00:19:58] [SPEAKER_01]: There's not discussions in the hallway or anything else.

[00:20:02] [SPEAKER_01]: You're seeing that level of transparency today even now with the AI companies.

[00:20:07] [SPEAKER_01]: And so just this week or maybe it was last week now, Microsoft wrote a transparency note

[00:20:13] [SPEAKER_01]: about Copilot.

[00:20:15] [SPEAKER_01]: And I've been cited in a couple of articles in the register about problems people are

[00:20:20] [SPEAKER_01]: seeing with Copilot.

[00:20:21] [SPEAKER_01]: But Microsoft said, hey, here's our transparency note.

[00:20:25] [SPEAKER_01]: Yesterday, OpenAI said, hey, here's how we test our systems.

[00:20:29] [SPEAKER_01]: Here's our model card for OpenAI, for GPT.

[00:20:35] [SPEAKER_01]: Companies can do the same.

[00:20:36] [SPEAKER_01]: They're not publishing model cards or transparency notes about the technology, but they're saying,

[00:20:41] [SPEAKER_01]: hey, here's how we're using the technology.

[00:20:43] [SPEAKER_01]: Here's how we're monitoring it.

[00:20:45] [SPEAKER_01]: Here's how we're going to operate.

[00:20:48] [SPEAKER_01]: And I think simply by writing it down and being transparent about that, then it has

[00:20:55] [SPEAKER_01]: positive implications, positive effects throughout then the operations of all these systems.

[00:21:05] [SPEAKER_00]: And I'm very conscious that throughout our conversations, they would be very forward-looking

[00:21:09] [SPEAKER_00]: and giving valuable takeaways for people listening.

[00:21:13] [SPEAKER_00]: And you've had a fantastic career.

[00:21:14] [SPEAKER_00]: But of course, none of us are able to achieve any degree of success without a little help

[00:21:18] [SPEAKER_00]: along the way.

[00:21:19] [SPEAKER_00]: Very often, there's someone that we could be grateful towards who maybe saw something

[00:21:23] [SPEAKER_00]: in us or invested a little time that just helped us get us where we are today.

[00:21:27] [SPEAKER_00]: Who would that person be for you or people?

[00:21:30] [SPEAKER_00]: And maybe we'll give them a little shout out to them.

[00:21:33] [SPEAKER_01]: Yeah, let me give you two people.

[00:21:35] [SPEAKER_01]: And I don't know if they're hearing this or not, but I'll just mention their names.

[00:21:40] [SPEAKER_01]: So when I was in graduate school way back in 1990, I had an offer to go working at an

[00:21:49] [SPEAKER_01]: aircraft company.

[00:21:51] [SPEAKER_01]: And I had a professor, a guy named Paul Kimerling, take me aside and say, you don't want to go

[00:21:57] [SPEAKER_01]: do that because after 30 years, you're going to be in charge of the left lug nuts on the

[00:22:02] [SPEAKER_01]: tires.

[00:22:02] [SPEAKER_01]: And there'll be somebody else in charge of the right lug nut.

[00:22:06] [SPEAKER_01]: Go explore the world.

[00:22:07] [SPEAKER_01]: Go take many jobs.

[00:22:10] [SPEAKER_01]: Go be intellectually curious.

[00:22:12] [SPEAKER_01]: Go try things.

[00:22:16] [SPEAKER_01]: Go wander through your career.

[00:22:18] [SPEAKER_01]: You're going to be so much happier.

[00:22:20] [SPEAKER_01]: So Paul was like a major influence in that.

[00:22:23] [SPEAKER_01]: And it was interesting for me, for my personality.

[00:22:26] [SPEAKER_01]: That was super interesting.

[00:22:27] [SPEAKER_01]: That's what actually brought me to this point.

[00:22:30] [SPEAKER_01]: The other thing was probably 10 years later, yeah, probably 10 years later, there was a

[00:22:36] [SPEAKER_01]: guy named Dave White.

[00:22:37] [SPEAKER_01]: I was working on DARPA-based systems.

[00:22:41] [SPEAKER_01]: So DARPA is a big research organization.

[00:22:44] [SPEAKER_01]: And he looked at me and said, you know, you're doing all this great research, but you're

[00:22:48] [SPEAKER_01]: actually a decent leader.

[00:22:50] [SPEAKER_01]: So why don't you start to lead some things?

[00:22:54] [SPEAKER_01]: And that was a little kick that I needed in my career.

[00:22:58] [SPEAKER_01]: And it's been great.

[00:23:00] [SPEAKER_01]: You know, I often say this to my teams, you know, I may forget some details of a project

[00:23:09] [SPEAKER_01]: here and there, but I always remember the people that we've worked on those programs together.

[00:23:15] [SPEAKER_01]: And, you know, throughout my career, I just remember those events.

[00:23:20] [SPEAKER_01]: And sometimes they're, you know, like those critical events, you know, like what are we

[00:23:25] [SPEAKER_01]: going to do?

[00:23:26] [SPEAKER_01]: COVID's hitting.

[00:23:27] [SPEAKER_01]: How are we going to support our clients?

[00:23:28] [SPEAKER_01]: How are we going to get people with their computers?

[00:23:31] [SPEAKER_01]: You know, whatever it happens to be.

[00:23:33] [SPEAKER_01]: But it's not about, you know, the technical part.

[00:23:36] [SPEAKER_01]: It's not about anything.

[00:23:38] [SPEAKER_01]: What I remember is the teams of people and the interactions with the people coming together.

[00:23:44] [SPEAKER_01]: And so from my perspective, yeah, there's lots of achievements that the teams have made,

[00:23:49] [SPEAKER_01]: but it's really those relationships and those interactions with people have been the most

[00:23:54] [SPEAKER_01]: rewarding part of this career in AI and security.

[00:23:59] [SPEAKER_00]: What a fantastic answer and a perfect moment to end on.

[00:24:02] [SPEAKER_00]: I think we often get carried away talking about the technology, the AI, but as you rightly

[00:24:07] [SPEAKER_00]: pointed out there, the role of people and everybody coming together, collaborating and

[00:24:11] [SPEAKER_00]: bringing those things to life.

[00:24:12] [SPEAKER_00]: So important.

[00:24:13] [SPEAKER_00]: And equally naming those people that have had a real impact on you and your life there.

[00:24:18] [SPEAKER_00]: They probably know you very well, but probably unaware of just how much you have impacted

[00:24:23] [SPEAKER_00]: their life, your life and your career.

[00:24:26] [SPEAKER_00]: So, so important that we recognize that.

[00:24:28] [SPEAKER_00]: And I genuinely hope that they get to hear that heartfelt message.

[00:24:32] [SPEAKER_00]: And for everyone listening, just wanting to find out more about security AI and maybe

[00:24:38] [SPEAKER_00]: connect with you or your team.

[00:24:40] [SPEAKER_00]: Where would you like to point everyone?

[00:24:41] [SPEAKER_01]: Well, obviously you can go to our website security.ai and that's security with an I.

[00:24:48] [SPEAKER_01]: So, and lots of information there.

[00:24:51] [SPEAKER_01]: There's a certification that you can take in terms of just learning about generative

[00:24:55] [SPEAKER_01]: AI.

[00:24:56] [SPEAKER_01]: I took it and learned a few things.

[00:24:58] [SPEAKER_01]: They get to do that.

[00:24:59] [SPEAKER_01]: You can connect with me on LinkedIn, Jack Berkowitz.

[00:25:04] [SPEAKER_01]: And happy to do that.

[00:25:06] [SPEAKER_01]: And, and look for us at a few conferences coming up.

[00:25:10] [SPEAKER_01]: You know, you can always find us at, you know, those, those normal data conferences or security

[00:25:15] [SPEAKER_01]: conferences.

[00:25:16] [SPEAKER_01]: For Europe, I'll be in Brussels at the IAPP conference in, in, in November.

[00:25:24] [SPEAKER_01]: And if there's any privacy and security professionals there, love to see it in Brussels.

[00:25:30] [SPEAKER_01]: I used to live in Brussels as well and love that city.

[00:25:34] [SPEAKER_00]: Awesome.

[00:25:35] [SPEAKER_00]: I'll connect with you on LinkedIn as well, because I am in the US a lot for tech conferences,

[00:25:40] [SPEAKER_00]: but I think my next to a VMware in Barcelona and Lenovo is in Seattle, I believe.

[00:25:46] [SPEAKER_00]: So hopefully our paths will cross in the near future too.

[00:25:50] [SPEAKER_00]: But just love chatting with you today about proven strategies for implementing AI risk

[00:25:54] [SPEAKER_00]: management frameworks, enhancing security through AI governance and better ways to evaluate

[00:26:01] [SPEAKER_00]: new AI capabilities, but bringing it to life with real world examples from some of those

[00:26:05] [SPEAKER_00]: conversations you've been having with clients and prospects that are all advancing data

[00:26:10] [SPEAKER_00]: and AI implementations.

[00:26:12] [SPEAKER_00]: I think everyone's on that same journey right now, but also for just wrapping it up with

[00:26:16] [SPEAKER_00]: a very human story of gratitude and the people you're grateful towards for putting you where

[00:26:20] [SPEAKER_00]: you are now.

[00:26:21] [SPEAKER_00]: But thanks for sharing that powerful story with me today.

[00:26:24] [SPEAKER_01]: Thank you.

[00:26:25] [SPEAKER_01]: It's really enjoyed the conversation.

[00:26:27] [SPEAKER_00]: I think Jack has given us valuable insights today into how organisations can navigate

[00:26:32] [SPEAKER_00]: the growing demands of AI innovation without sacrificing security or compliance.

[00:26:41] [SPEAKER_00]: And from the importance of a strong AI governance framework to the proactive measures needed

[00:26:45] [SPEAKER_00]: to protect data, I think it's clear that the road ahead requires both strategic foresight

[00:26:52] [SPEAKER_00]: and operational precision.

[00:26:53] [SPEAKER_00]: But the big question is how will your business balance innovation with risk management?

[00:27:00] [SPEAKER_00]: And what steps will you take to ensure AI transparency and trust across your organisation?

[00:27:07] [SPEAKER_00]: I think as Jack has shown today, the answers lie in understanding not only the technology,

[00:27:13] [SPEAKER_00]: but also the ethical and regulatory landscape shaping the future of AI.

[00:27:18] [SPEAKER_00]: But as always, you know the drill.

[00:27:19] [SPEAKER_00]: Let me know your thoughts.

[00:27:20] [SPEAKER_00]: Tech blog writer outlook.com.

[00:27:22] [SPEAKER_00]: Send me a quick message on LinkedIn, Instagram, Twitter, just at Neil C Hughes.

[00:27:26] [SPEAKER_00]: And we'll keep this conversation going.

[00:27:29] [SPEAKER_00]: But that's it quitting time for me.

[00:27:30] [SPEAKER_00]: Time for me to knock off for the day, but I'll return again tomorrow.

[00:27:34] [SPEAKER_00]: Thanks for listening today and hopefully I will speak with you all again tomorrow.

[00:27:37] [SPEAKER_00]: Bye for now.