РАСШИФРОВКА 🔗
WEBVTT
00:00:03.759 --> 00:00:08.000
You are now listening to the Secure AF Podcast.
00:00:12.080 --> 00:00:14.400
Welcome to the Secure AF Podcast.
00:00:14.480 --> 00:00:21.679
Today we're going to be talking about AI, AI governance, and all kinds of topics that everyone wants to know, but they're kind of afraid to ask.
00:00:21.839 --> 00:00:24.800
So, Richard Call, thank you so much for joining me.
00:00:24.879 --> 00:00:27.039
Um, I'm really excited for you being here.
00:00:27.280 --> 00:00:30.160
And uh we've we've got some fun topics here.
00:00:30.239 --> 00:00:35.359
And I know AI is all over the news and it's all over every conference you go to.
00:00:35.679 --> 00:00:38.880
But today we're going to be kind of talking about AI governance.
00:00:38.960 --> 00:00:44.479
And this is something that a few people are talking about, but we're people aren't going into a lot of details.
00:00:44.640 --> 00:00:46.799
AI is still just a little new.
00:00:47.119 --> 00:00:49.280
So, first, tell us who you are.
00:00:49.359 --> 00:00:51.359
You know, what do you do, who you are, where are you from?
00:00:51.679 --> 00:00:52.719
Thanks so much, Jonathan.
00:00:53.039 --> 00:00:54.320
So happy to be here today.
00:00:54.479 --> 00:00:55.359
I'm Richard.
00:00:55.439 --> 00:01:01.280
I'm the CEO and founder of Compliance with Awai, C-O-M-P-L-Y-A-N-C-E.
00:01:01.520 --> 00:01:06.159
It's a bit of a play on words because we are focused on the reasons behind compliance.
00:01:06.400 --> 00:01:16.400
We see compliance as not just a checkbox exercise, but we consider that really strong security compliance is a byproduct of really good security practices.
00:01:16.640 --> 00:01:31.120
And we use AI and automation to essentially support our GRC teams and security teams who we work with to change the work that they do from the more manual work to the more judgment-driven work using our purpose-built agents.
00:01:32.079 --> 00:01:32.719
Very nice.
00:01:32.879 --> 00:01:33.120
Okay.
00:01:33.280 --> 00:01:35.439
So I gotta ask, how did you get into this?
00:01:35.599 --> 00:01:38.480
How did you get into this sort of work?
00:01:38.879 --> 00:01:44.319
Yeah, so I have a very um sort of fun and very personal path to it.
00:01:44.480 --> 00:01:53.359
Uh, first things first is that I actually was doing a lot of public sector consulting on things like uh policies for uh the tech industry.
00:01:53.519 --> 00:01:56.959
So I started my career at McKinsey, where I did public sector work only.
00:01:57.120 --> 00:02:00.079
And then I worked for the state of Virginia in the US.
00:02:00.319 --> 00:02:04.959
And we actually um actually was leading the growth of the tech sector for the state.
00:02:05.120 --> 00:02:20.400
And that included sitting on committees for things like drone privacy regulation and other sorts of privacy and security regulations that basically help to balance innovation with the protection of uh data and people.
00:02:20.879 --> 00:02:24.479
And
Load full transcript
so I kind of came at it from the government side.
00:02:24.719 --> 00:02:33.840
Then I um then I actually, you know, led GRC and security uh for the beginning of my tenure as chief strategy officer at a legal AI company.
00:02:33.919 --> 00:02:36.080
So I experienced some of that firsthand.
00:02:36.319 --> 00:02:39.199
Uh, but in short, I'm just a data privacy nut.
00:02:39.280 --> 00:02:41.919
I really like to keep my data secure.
00:02:42.159 --> 00:02:49.039
Um, my friend, one of my really close friends, was part of that Equifax breach that happened, I don't know, 10 years ago now.
00:02:49.360 --> 00:02:54.639
Do you remember that was that first breach that made everyone kind of aware of data breaches?
00:02:54.800 --> 00:02:55.280
Yep.
00:02:55.439 --> 00:02:55.840
Yep.
00:02:56.080 --> 00:02:58.879
And they spent years recovering from that breach.
00:02:59.039 --> 00:03:02.400
And it was because of things that they couldn't control.
00:03:02.719 --> 00:03:14.960
And I started coming at it from the perspective of saying, hey, in order to protect, you know, ourselves, our parents, our friends, our coworkers, we need to actually help secure the enterprises who hold the world's data.
00:03:15.120 --> 00:03:16.080
And that's what we do.
00:03:16.400 --> 00:03:17.199
Well, I love that.
00:03:17.360 --> 00:03:19.439
I'm a I'm a data privacy guy as well.
00:03:19.599 --> 00:03:36.800
Um, and that's kind of they they they all think I'm weird here at the cybersecurity company of alias because I love doing the compliance, I like doing contracts, I like doing the the the vendor management and all the different things you have to do to really make sure that your data is safe.
00:03:37.199 --> 00:03:43.360
Um so this isn't part of kind of the things that we talked about, but I have a question about this.
00:03:43.520 --> 00:03:59.039
So when we're talking about data, this AI conversation that we're having, I've always approached it with data first, you know, make sure your data storage policy, your data communication policy, all these different things that you have within the organization, make sure that's strong.
00:03:59.120 --> 00:04:05.039
And AI is just another tool, but that's not how people are using it, and that's not really how it's being deployed.
00:04:05.360 --> 00:04:15.680
What are you seeing as kind of your top data concern as it relates to AI and more specifically, organizations using AI?
00:04:16.240 --> 00:04:16.879
Yeah.
00:04:17.279 --> 00:04:22.720
I actually think that the most pressing concern right now is third-party AI.
00:04:22.959 --> 00:04:36.399
I think a lot of companies have a slew of vendors who have been adding AI into their products without necessarily that flagging additional reviews or security checks on the company side.
00:04:36.639 --> 00:04:50.079
I think that from a shadow IT perspective, we have employees everywhere who are putting confidential information from the company into AI tools that they are testing out, want to use, and so on.
00:04:50.319 --> 00:04:54.879
On the third party side, there's risks that are really hard to see.
00:04:55.199 --> 00:05:03.279
And I think that there are certainly risks on the internal side, you know, the use of AI internally and products, but there's so much more visibility.
00:05:03.439 --> 00:05:07.759
And I think that lack of visibility is actually what drives some of the biggest risk.
00:05:07.920 --> 00:05:12.879
So I'd have to put the biggest concern right now for me on the third-party AI side.
00:05:13.279 --> 00:05:14.240
Yeah, I agree.
00:05:14.319 --> 00:05:22.800
And we're seeing a lot with issues in terms of like bring your own device as people are using personal AIs to do work.
00:05:23.279 --> 00:05:37.600
Uh, we had this happen in a client of mine where they had doctors who were putting in patient data on their phones on an AI because the organization restricted AI usage and they figured out a way around it.
00:05:37.839 --> 00:05:43.759
So it is absolutely a concern, um, which does lead into this AI governance thing.
00:05:43.920 --> 00:05:55.839
And when we talk about governance, um talking about GRC specifically, um, but what do you think, where do you think we're at right now?
00:05:56.000 --> 00:06:05.839
You know, if you had to kind of see the the industry and the different communities and different you know infrastructures that we have, where do you think AI governance stands?
00:06:05.920 --> 00:06:08.480
Are are we just now getting into it?
00:06:08.639 --> 00:06:11.040
Have we tried to do it and it hasn't been working?
00:06:11.199 --> 00:06:13.199
Where where do you feel like we are?
00:06:14.319 --> 00:06:22.240
So last week, actually, I was in Phoenix at a GRC summit with, I don't know, a hundred plus GRC professionals.
00:06:22.399 --> 00:06:25.759
And I was leading a discussion on AI governance.
00:06:25.920 --> 00:06:35.360
The first question that was on everyone's minds is basically should AI governance be its own program or should it be melded into the GRC team?
00:06:36.240 --> 00:06:39.120
And I think that's a really, really important question.
00:06:39.279 --> 00:06:42.079
I'd love to share some of the takeaways from that discussion here.
00:06:42.319 --> 00:06:42.879
Yeah, of course.
00:06:43.279 --> 00:06:43.680
Jonathan.
00:06:43.759 --> 00:06:58.720
I think, you know, we had folks in the room from um Bank of America, CBRE, you know, leading household brands, all kind of chiming in, Alaska Airlines, many others, on, you know, how they're approaching it.
00:06:58.959 --> 00:07:06.560
And the the summary that we've seen from our clients and the folks in that room is that AI governance is still at its infancy.
00:07:06.959 --> 00:07:09.680
People are struggling with how to start.
00:07:09.920 --> 00:07:21.199
The conversations that I hear about it feel loose compared to conversations that we have with clients about their data security and their information security more broadly.
00:07:21.600 --> 00:07:27.199
Folks can speak much more closely to things like privacy controls than they can speak to AI controls.
00:07:27.439 --> 00:07:28.399
And that makes sense.
00:07:28.560 --> 00:07:33.839
We're in the infancy of learning about this as a topic for across, you know, society right now.
00:07:34.319 --> 00:07:37.040
But I think that it's tough when that's happening.
00:07:37.199 --> 00:07:40.720
And so naturally the first question becomes well, how do I staff this up?
00:07:40.800 --> 00:07:42.079
Like where does this go?
00:07:42.240 --> 00:07:43.360
Where does this live?
00:07:43.680 --> 00:08:00.000
What I shared with the group last week and and I want to share now is that I think the question of whether AI governance needs to be its own program or should be melded into the GRC program should be based on the capacity of your current GRC team and function.
00:08:00.319 --> 00:08:00.959
Right.
00:08:01.360 --> 00:08:16.560
And to me, there needs to be a rapid assessment, almost like a maturity assessment and an honest one of how each of the internal GRC functions can actually cope with the need for AI governance.
00:08:16.720 --> 00:08:25.439
So, what I mean by that, there's gonna be some sort of internal function for compliance, compliance within I, that is, and you know, getting through your audits.
00:08:25.600 --> 00:08:33.200
Okay, well, can they cope with ISO 42001 controls or EU AI Act controls or you know, NIST AI RMF?
00:08:33.360 --> 00:08:35.039
Do they have the capacity to do that?
00:08:35.200 --> 00:08:39.279
That's kind of the first, the first subfunction that needs to be assessed.
00:08:39.519 --> 00:08:49.759
Secondly, your third party team is your third-party risk team and your third-party diligence team, are they the ones who are going to be, you know, reviewing these AI vendors?
00:08:49.840 --> 00:08:51.120
Do they have the capacity for it?
00:08:51.279 --> 00:08:53.840
Do they have the knowledge to ask the right questions?
00:08:54.080 --> 00:08:57.039
And then third, what about the risk team?
00:08:57.200 --> 00:09:03.919
Do you have a risk register and a risk assessment functionality right now that is alive, as I like to say?
00:09:04.080 --> 00:09:07.039
Is it just a stale list or is it actually continuously growing?
00:09:07.120 --> 00:09:12.960
And do you have that capacity internally to go and seek, okay, what are the actual AI risks that we need to deal with?
00:09:13.120 --> 00:09:19.679
And getting those on paper in a way that's quantified and, you know, able to be discussed at a board.
00:09:20.000 --> 00:09:21.519
Those are the primary three.
00:09:21.679 --> 00:09:26.720
There's also the piece around just having some sort of like knowledge internally to write policies.
00:09:26.879 --> 00:09:41.919
But I would say that if we can do a rapid maturity assessment on those three subfunctions of GRC, then you're basically getting a quick answer to, hey, do I need to have a separate team for AI governance or should I have my own?
00:09:42.080 --> 00:09:44.399
Because AI governance is moving quickly.
00:09:44.559 --> 00:09:48.559
And so if the team can't handle it, it's not a question that it needs to get done.
00:09:48.799 --> 00:09:51.360
So in that case, it needs to be a separate program.
00:09:51.759 --> 00:09:54.159
And either way, we support our clients with this all the time.
00:09:54.240 --> 00:09:55.519
So happy to go more in depth.
00:09:55.600 --> 00:10:01.039
But curious, Jonathan, does that kind of align to what you would also see and share with folks?
00:10:01.519 --> 00:10:02.639
Yeah, it is.
00:10:02.879 --> 00:10:08.399
So I'm I'm teaching an AI governance course at a conference here in a couple of weeks.
00:10:08.639 --> 00:10:14.080
And I'll tell you, they uh they contacted me and they said it sold out three or four times.
00:10:14.159 --> 00:10:16.080
They kept moving it to a different room.
00:10:16.320 --> 00:10:18.240
People are just hungry for it.
00:10:18.320 --> 00:10:21.840
They're seeing these risks, but they they don't know how to handle it.
00:10:22.159 --> 00:10:30.480
And, you know, a lot of the things that I ask when I go and I work with organizations is do you just like you said, do you have a risk registry?
00:10:30.639 --> 00:10:32.240
Do you understand what that means?
00:10:32.399 --> 00:10:35.039
Do you have your governance processes right now?
00:10:35.440 --> 00:10:37.120
And a lot of times they don't.
00:10:37.279 --> 00:10:57.840
Um, you know, they they don't have good compliance requirements, even for the regular cybersecurity and regular privacy things, adding the complication of AI, which generally speaking, if you've got good cyber and good privacy and good vendor management, you can have a pretty good starting point for your AI governance.
00:10:58.159 --> 00:11:00.799
But there's a lot of organizations that just don't have that.
00:11:00.960 --> 00:11:21.200
Um, and now my my my question to you is knowing that these organizations don't have that, and the the people, especially that might be listening to this podcast or might be seeing you at a conference, what do you tell them that they can do to really get that support from their leadership to begin that AI governance?
00:11:21.279 --> 00:11:26.320
How do they make it a priority or uh put importance on it for the organization?
00:11:27.120 --> 00:11:29.120
You know, it it's it's interesting, Jonathan.
00:11:29.200 --> 00:11:38.960
I I haven't seen people have to make it that much of a case because I think that it feels at least to our clients, again, please tell me if you're seeing something else.
00:11:39.200 --> 00:11:48.320
It feels like AI has become a topic that has permeated, you know, so quickly the minds of board members when it comes to risk.
00:11:48.559 --> 00:11:58.399
I actually think that AI has brought GRC to the boardroom in a way that no other really force has been able to do, because we all understand it.
00:11:58.639
--> 00:12:05.840
We all are exposed to Claude to Chat GPT, and we kind of intuitively get the fact that we're giving a lot of data.
00:12:06.000 --> 00:12:08.240
And what if our employees are doing that?
00:12:08.399 --> 00:12:15.440
And so I think it actually has become maybe more of the next step, which is like, okay, what do you need?
00:12:15.600 --> 00:12:17.200
And how much money do you need?
00:12:17.360 --> 00:12:22.240
And when can we see you know confidence in our own policies and security?
00:12:22.399 --> 00:12:24.480
And I think folks are real struggling with that.
00:12:24.639 --> 00:12:26.240
But let me just turn the question back to you.
00:12:26.399 --> 00:12:30.639
Do you feel like people are still having to make the case from the folks you're speaking with?
00:12:31.600 --> 00:12:32.879
I some, yes.
00:12:33.039 --> 00:12:43.200
I mean, I think that there is some aspects where people are well, I mean, honestly, we still see we still see a lot of organizations that have just outright forbidden AI.
00:12:43.360 --> 00:12:46.240
You know, they're not wanting it used across the board.
00:12:46.399 --> 00:12:54.879
Um, and then we have some that are just very much in the infancy of understanding how AI could be used within the organization.
00:12:55.440 --> 00:12:57.360
So I think that there is a lot.
00:12:57.519 --> 00:13:08.960
Um, I would hope, you know, I think if you had asked me this question a year ago, um, I would say that we're we're kind of we're we're picking up steam, we're picking up that momentum.
00:13:09.600 --> 00:13:12.000
Um, I don't think that we're there right now.
00:13:12.159 --> 00:13:23.039
I think right now for many organizations, small to medium business, even some large organizations, you know, the people that I talk to at conferences when I'm speaking and things, um, and the questions I'm getting.
00:13:23.200 --> 00:13:35.600
You know, I've built a fair amount of uh tabletops and things that have been released out there, and people are running through them and they keep contacting me saying, We don't know how to think about this.
00:13:35.679 --> 00:13:50.639
You know, we did your tabletop and we just we we we didn't understand what the risks were, you know, the the life safety, the the financial risk, and the the uh reputational risks, you know, that harm that could happen.
00:13:50.879 --> 00:13:52.639
So I think that people are growing.
00:13:52.720 --> 00:13:54.559
Now I think that there is a divide.
00:13:54.639 --> 00:13:55.600
You know, I think you're right.
00:13:55.679 --> 00:13:59.919
From a boardroom perspective, there is a lot of people that are talking about AI.
00:14:00.159 --> 00:14:08.960
They want AI, they, they, they know there's value there, whether it's efficiency, whether it's new technology, you know, whatever it may be.
00:14:09.279 --> 00:14:15.519
And then you've got this middle ground of the the worker within the organization there kind of in the dark.
00:14:15.840 --> 00:14:21.919
And I don't want to say that in a bad way, because that's not what I mean, but they don't use it every day or they don't know how to use it.
00:14:22.080 --> 00:14:28.399
But then you have the IT teams who are afraid of it because they don't know how to manage it, they don't know how to control it.
00:14:28.639 --> 00:14:47.200
So I think that there's, and that's where I've come in, you know, what I'm trying to do is build that, you know, that communication, you know, from maybe a CISO, maybe a CIO or maybe a technology team to come in and say, okay, how do we help you as the business engage this new technology?
00:14:47.360 --> 00:14:48.399
Because that's really all it is.
00:14:48.480 --> 00:14:56.559
I mean, it's the equivalent of cloud storage back from 15, 20 years ago, you know, when we were talking about box and drop box and things.
00:14:56.720 --> 00:14:58.240
Everyone was freaking out about it.
00:14:58.399 --> 00:15:00.639
Now you really can't live without it.
00:15:00.799 --> 00:15:04.159
You go back 30, 40 years ago, it's the internet.
00:15:04.639 --> 00:15:12.720
You know, we've had these things that have happened to the organization that and the organizations had to evolve to utilize these technologies.
00:15:13.039 --> 00:15:16.480
AI is another technology, um, and it's fantastic.
00:15:16.559 --> 00:15:19.919
It's it's powerful, you could become very efficient in it.
00:15:20.399 --> 00:15:22.399
There are risks for sure.
00:15:22.720 --> 00:15:28.240
But I think the biggest thing is getting that conversation started because you're right, the board is wanting it.
00:15:28.639 --> 00:15:36.000
IT maybe is afraid of it, and everyone in the middle is just like, eh, just I just want to do my job, you know.
00:15:36.159 --> 00:15:54.799
So getting the conversation started across the board is really important, at least in my world, it's really important because if you don't start asking questions to figure out how it's going to be used, why it's going to be used, and how we can make it safe, um, then people are just kind of do off doing their own thing.
00:15:55.120 --> 00:15:55.440
Yeah.
00:15:55.600 --> 00:15:55.919
Yeah.
00:15:56.080 --> 00:15:57.360
No, I totally agree.
00:15:57.519 --> 00:16:00.559
And so then the question becomes okay, so what are those first steps?
00:16:00.720 --> 00:16:07.519
And I think that we can go back to that idea of internal versus external AI risk.
00:16:07.600 --> 00:16:09.600
And I really anchor on that with our clients.
00:16:10.000 --> 00:16:12.559
Internally, you have to get your policies in order.
00:16:12.720 --> 00:16:17.440
You have to set up internal controls to basically make sure that those policies are enforced.
00:16:17.679 --> 00:16:22.480
You have to actually make sure that your risk register is reflecting the risks of AI.
00:16:22.559 --> 00:16:26.960
And you have to make sure that your internal processes for external vendors are in place.
00:16:27.120 --> 00:16:29.200
That's really the summation of what it is.
00:16:29.360 --> 00:16:36.559
Of course, now if you're creating AI products, there's a host of other controls, which is a separate, separate conversation.
00:16:36.799 --> 00:16:43.440
But that is the, those are the beginnings of an AI governance program that are at least first steps that everyone is familiar with.
00:16:43.679 --> 00:16:48.000
Then we need to talk about the external side, which is the third-party risk management.
00:16:49.039 --> 00:16:51.200
Let's go back to the internal real quick.
00:16:51.440 --> 00:17:03.919
If I was a brand new GRC person, you know, I've got a couple of years under my belt, and now I'm now I'm sitting there and someone comes to me and says, All right, we need AI governance.
00:17:04.319 --> 00:17:06.799
What is the first thing that person should do?
00:17:06.960 --> 00:17:13.759
You know, they they're they're a little experienced, they've they've got an understanding, but they don't know anything about this, the AI governance stuff.
00:17:13.920 --> 00:17:17.440
What would you tell them that this is the very first thing you need to do?
00:17:17.759 --> 00:17:18.400
Yeah.
00:17:18.880 --> 00:17:22.319
I really think that I would go to where the risk is highest.
00:17:22.400 --> 00:17:31.519
And that means I would actually start to inform myself of what are the risks from my third parties to my organization.
00:17:31.759 --> 00:17:33.039
I can just share them now.
00:17:33.200 --> 00:17:49.839
You you can have third parties who are using AI subprocessors that are not sending the data to the um, or sorry, that are sending your customer data to their subprocessors without any real oversight, knowledge.
00:17:49.920 --> 00:17:58.559
So your confidential data could be getting passed through back to those uh, you know, whether it's the LLMs or whether it's some other providers, you don't really need to know that.
00:17:58.799 --> 00:18:06.000
Usually subprocessor notifications just kind of pass us by, as we kind of know, once you've, you know, confirmed a vendor, they just kind of happen in the background.
00:18:06.240 --> 00:18:12.559
And so I think that when when we're thinking about it, I really do think we have to go to where the risk is highest.
00:18:12.720 --> 00:18:30.720
So what I would do is then either partner with someone, you know, who has out-of-the-box questionnaires, like we do, and I know others do as well, where we've already created those assessments for the third parties and you can send them out all in one go and say, look, we need to do a rapid, you know, AI governance assessment on all of our vendors.
00:18:30.880 --> 00:18:32.720
Here are the seven questions we need you to answer.
00:18:32.880 --> 00:18:35.519
If you say yes, maybe it becomes 15 at most, you know.
00:18:35.759 --> 00:18:43.680
And then we have an AI agent that does an automatic review of those answers and flags gaps in like human layman's English.
00:18:44.000 --> 00:18:51.920
And
so what happens is that you don't need a ton of knowledge going into it, but you do need to say, okay, this is what I'm gonna prioritize.
00:18:52.000 --> 00:18:54.000
I'm gonna go and do an assessment of my vendors.
00:18:54.079 --> 00:18:58.079
I'm gonna partner with someone who knows the questions to ask, how to interpret the answers.
00:18:58.240 --> 00:19:00.319
And I'm gonna have a list of findings now.
00:19:00.480 --> 00:19:15.359
And then I need to inform myself of whether those findings need to be raised as risks and this vendor needs to be disqualified, or if this vendor just needs, you know, we need to have conversations that say, hey, we need you to adapt this, or we need to add an addendum to our contract around AI privacy, or so on or so on.
00:19:15.519 --> 00:19:17.440
And that's literally the work that we do.
00:19:18.319 --> 00:19:35.119
So I'll I'll give a story for our listeners, and I think you'll you'll find this maybe not amusing, maybe scary, but had a client who uh they contacted me and I'm their VC so and they said, Hey, we have a new tool that's in this.
00:19:35.359 --> 00:19:41.519
Um we have a new AI functionality in this tool that we we want to buy.
00:19:41.920 --> 00:19:43.440
I was like, okay, great.
00:19:43.599 --> 00:19:46.559
Um, I need to see your, you know, the the MSA.
00:19:46.799 --> 00:19:49.519
I want to see the BAA, I want to see all these pieces.
00:19:49.759 --> 00:19:52.720
And uh they got me on the phone with them, you know, with the vendor.
00:19:52.880 --> 00:19:56.000
I was like, hey, just I have some questions, just basic stuff.
00:19:56.160 --> 00:19:59.359
And I asked them what AI they were using.
00:19:59.680 --> 00:20:04.319
And they were like, well, you know, a couple of years ago we started building and this and that.
00:20:04.640 --> 00:20:07.680
Like, okay, that that doesn't answer my question.
00:20:07.759 --> 00:20:10.079
I need to know what AI are you using right now.
00:20:10.319 --> 00:20:12.160
Well, they're using ChatGPT.
00:20:12.319 --> 00:20:14.559
They had a wrapper for ChatGPT.
00:20:14.799 --> 00:20:20.640
And I'm like, um, you can't use it for this industry because it's not HIPAA compliant.
00:20:20.799 --> 00:20:21.279
Yep.
00:20:21.599 --> 00:20:27.759
Um, and they're like, oh no, no, no, you have a BAA with us, so everything we do down the street.
00:20:27.839 --> 00:20:29.839
It's like, no, that's not the way this works.
00:20:30.000 --> 00:20:33.279
So I worked with the client and they understood it.
00:20:33.440 --> 00:20:40.960
I mean, they understood it from a BAA point of view, but they didn't understand that subprocessor and that sub-subprocessor and on down the line.
00:20:41.680 --> 00:20:48.160
And so we ended up, we had to turn that functionality off because they it was absolutely going to see PHI.
00:20:48.640 --> 00:20:50.559
They could not get us a BAA.
00:20:51.359 --> 00:21:04.960
And so the salesman, and here's the thing, and this is the warning to the audience the salesman who was selling this product, he thought he knew he was completely covered and everything was fine and protected.
00:21:05.279 --> 00:21:14.000
So, back to what you said earlier is that biggest risk in that third-party um management side of are we actually protected?
00:21:14.160 --> 00:21:16.720
And you're right, you have to ask the right questions.
00:21:16.799 --> 00:21:23.759
And sometimes you have to ask the right people because the first answer you get from the salesperson, and I'm not knocking salespeople, I'm really not.
00:21:23.920 --> 00:21:26.000
I've got a lot of salespeople that are really good friends.
00:21:26.079 --> 00:21:27.920
They just they may not know.
00:21:28.160 --> 00:21:31.119
And for most industries, they're fine.
00:21:31.440 --> 00:21:34.720
For hospitals, though, that's very different.
00:21:35.039 --> 00:21:43.920
So I I it it warms my heart that the first thing you said is third party management, because I I beat that drum all the time with my clients, and it's really important.
00:21:44.640 --> 00:21:49.759
Um, to that end, um, thinking about again that brand new GRC person.
00:21:51.039 --> 00:21:56.240
The hardest thing really is when you find problems getting change.
00:21:56.480 --> 00:21:59.759
Um, I wanted to kind of get your thoughts on let's say that.
00:22:00.160 --> 00:22:07.680
GRC person does find an issue with a vendor and a problem with the data and the data movements through the data flow diagram, whatever, whatever it may be.
00:22:08.240 --> 00:22:18.880
How would you encourage them to take that next level, the next step of, okay, now I've got to tell the organization that this is a risk and that we need to make a change?
00:22:19.039 --> 00:22:23.440
Because honestly, that's really scary for a lot of people saying, hey, I found a problem.
00:22:23.519 --> 00:22:24.880
We need to change this.
00:22:25.359 --> 00:22:25.839
Yeah.
00:22:26.160 --> 00:22:36.799
To be very honest, I would quantify the risk and I would rank order it amongst other risks that the board or the business already knows and understands as risks.
00:22:37.039 --> 00:22:47.200
I actually think it's a good litmus test for the GRC person as well, because I think it forces a level of, again, ranking and prioritization to these problems.
00:22:47.359 --> 00:22:59.039
And if you can go to the team and say, hey, remember that thing we discussed or that thing that matters to you, you know, give the context and say this risk has potentially higher exposure for our organization.
00:22:59.279 --> 00:23:00.960
Now I know folks don't know where to start.
00:23:01.039 --> 00:23:04.559
And again, I think it's something we've been discussing with our clients a lot.
00:23:04.799 --> 00:23:08.559
We have like, as do many others, I think there's like the fair model.
00:23:08.640 --> 00:23:13.039
There's, you know, Monty, um, oh my goodness, Monte Carlo analysis.
00:23:13.279 --> 00:23:14.160
Did I just get that wrong?
00:23:14.240 --> 00:23:16.079
I I can't believe I just blanked on that.
00:23:16.400 --> 00:23:18.400
Monte Carlo risk analysis.
00:23:18.720 --> 00:23:20.319
This sounds right now that's in my head.
00:23:20.559 --> 00:23:22.960
I believe I just said it and I was like, wait a second.
00:23:23.039 --> 00:23:25.119
I was gonna say Monty Python out loud.
00:23:25.200 --> 00:23:26.559
So I'm glad I didn't say that.
00:23:29.359 --> 00:23:30.559
Anyway, yes.
00:23:30.720 --> 00:23:34.240
Um, so you know, there's all these ways to analyze risks.
00:23:34.400 --> 00:23:37.279
I just heard yesterday that folks are using credit scores as well.
00:23:37.440 --> 00:23:44.960
So um there's these conversations that are happening, but you know, we've actually made a risk quantification agent in our platform.
00:23:45.119 --> 00:23:47.039
It's getting released next next month.
00:23:47.200 --> 00:23:57.279
And it's because of this, it's literally to solve this use case to say, how do you start to rank order and prioritize your risks when everything feels so important right now?
00:23:57.519 --> 00:24:00.799
And how do you then do the narrative storytelling, right?
00:24:00.960 --> 00:24:07.920
Because it's so hard to be in that position, to go to the board as a security leader, as a GRC leader.
00:24:08.160 --> 00:24:13.680
You basically have the board expecting confidence, expecting answers, expecting visibility.
00:24:13.920 --> 00:24:20.400
And you're kind of like, look, my first lines of defense have other jobs that they're trying to do.
00:24:20.559 --> 00:24:23.359
They're not raising to me these risks proactively.
00:24:23.519 --> 00:24:27.200
I'm out here kind of trying to get a complete picture all the time.
00:24:27.359 --> 00:24:30.480
And so you have to enter that room and it sucks.
00:24:30.640 --> 00:24:31.519
It really sucks.
00:24:31.680 --> 00:24:34.480
And I think that the feeling of it doesn't feel good.
00:24:34.640 --> 00:24:40.319
And so, how do we arm and equip our clients to go ahead and actually feel better in that room?
00:24:40.559 --> 00:24:46.880
One of the ways is through, honestly, you know, instant risk quantification, which gives you risk prioritization.
00:24:47.039 --> 00:24:55.759
And then you can go in the room and tell a story and a narrative that's compelling and say, look, we talked about this last year, or you know, we're all aware of risks like this.
00:24:56.000 --> 00:25:01.839
Okay, so this AI risk has a has a higher likelihood and maybe lower impact or something along those lines.
00:25:02.000 --> 00:25:04.240
Our exposure is this many millions of dollars.
00:25:04.400 --> 00:25:08.240
And of course, the likelihood is lower, but this is how we're seeing it play out.
00:25:08.319 --> 00:25:10.319
And this is the research that backs that up.
00:25:10.400 --> 00:25:16.160
Um, and so that's actually the conversations that we're uh getting to now with our clients.
00:25:16.240 --> 00:25:26.319
And um, it's it's a pretty, it's a pretty amazing place to be, but it's also one that really puts into perspective where we are from a security, you know, position in the world right now.
00:25:26.480 --> 00:25:28.000
It's it's hard to be a CISO.
00:25:29.599 --> 00:25:30.160
It is.
00:25:30.319 --> 00:25:36.160
I actually uh I I was I was making some I do a lot of tabletops.
00:25:36.240 --> 00:25:43.440
Um I like I like working with organizations and drilling their procedures, their policies or procedures or playbooks.
00:25:43.759 --> 00:25:50.640
Um and an organization reached out to me and says, Hey, we would like you to build some of this for AI because people they want it.
00:25:50.720 --> 00:25:52.720
They're they're they're hungry for it.
00:25:53.119 --> 00:25:59.680
So I did it, and at the end of the recordings, you know, it was, I don't know, eight hours of recordings, whatever it was.
00:25:59.920 --> 00:26:06.720
They says, Hey, we would like to just interview you just as a as a content provider, you know, a SME.
00:26:06.880 --> 00:26:08.480
I was like, Oh, okay, sure.
00:26:08.960 --> 00:26:18.319
Well, I uh uh so we get on the call and we're recording kind of like this, and I made the comment that it's like, you know, sometimes this job sucks.
00:26:18.559 --> 00:26:20.799
And that's why we engage each other.
00:26:20.880 --> 00:26:24.240
That's why this this collaboration of professionals is so important.
00:26:24.480 --> 00:26:31.839
Because if I if I don't know how to do something, I'm gonna ask someone else who has done it, and I'm gonna get their expertise and get their experience.
00:26:32.079 --> 00:26:37.119
But the fact that I said this job sucks sometimes, everyone was silent.
00:26:37.279 --> 00:26:39.440
I was like, Oh, well, you guys can cut that out.
00:26:39.519 --> 00:26:40.079
That's fine.
00:26:40.319 --> 00:26:50.160
And uh I got a call about a month later, and they're like, We just saw that you said this on a live thing.
00:26:50.319 --> 00:26:52.319
We we need you on this.
00:26:52.480 --> 00:26:59.039
It was a webinar with several other people, but it was having that conversation that, yeah, this does suck sometimes.
00:26:59.200 --> 00:27:00.400
And it's hard.
00:27:00.559 --> 00:27:09.599
And when you take that brand new GRC professional, it's really hard for them to be able to stand up in front of someone and say, hey, this is a big risk.
00:27:09.680 --> 00:27:14.000
You know, I know you guys want this, but we have to talk about this.
00:27:14.240 --> 00:27:23.599
I uh I built a scoring system for a couple of my clients a while back, specifically for that risk matrix, you know, the standard, you know, impact versus likelihood.
00:27:23.680 --> 00:27:25.440
And I wanted to add to that.
00:27:26.079 --> 00:27:31.440
And uh and I was I started it with what kind of data are we dealing with?
00:27:31.599 --> 00:27:36.480
You know, if it's compliance data or in scope data, that has a higher, you know, weight to it.
00:27:36.640
--> 00:27:39.920
Because if you're writing letters, okay, fine, I don't care.
00:27:40.000 --> 00:27:43.839
But if you're writing letters with PII in it, uh I'm gonna care about that.
00:27:44.000 --> 00:27:51.599
Um, but then I I I put a weight on there for um uh what is the AI doing?
00:27:51.759 --> 00:27:52.559
Is it an agent?
00:27:52.640 --> 00:27:53.759
Is it an LLM?
00:27:53.920 --> 00:27:56.319
Is it you know a machine learning type?
00:27:57.119 --> 00:28:05.279
And then you had it was just one through five, you know, very basic, one through five on data, one through five on this, one through five on impact, one through five on likelihood.
00:28:05.599 --> 00:28:16.319
But what was interesting to me is when I had organizations start working through it from a leadership perspective, they very quickly understood it.
00:28:16.480 --> 00:28:20.240
And I mean, this is just kind of a fun thing for you and I.
00:28:21.839 --> 00:28:30.400
Because you've got four numbers and each one one through five, high, you know, higher the number, the higher the impact, or you know, the more sense of the data.
00:28:30.559 --> 00:28:32.960
That means you get really big numbers in the end.
00:28:33.279 --> 00:28:39.759
You know, when you say, Oh, it's five times five times five times five, you got a really big number.
00:28:41.200 --> 00:28:48.160
And the weird thing was when I was in the room, everyone figured out immediately big numbers are bad.
00:28:48.400 --> 00:28:50.240
Big numbers mean more risk.
00:28:50.480 --> 00:28:55.200
Big numbers means we need to do a little bit extra to make sure we're protected.
00:28:55.920 --> 00:29:20.000
And it it the the idea of I'm gonna say manipulating the the leadership or manipulating the the people you're working with, the stakeholders we talk about this a lot, but that little piece I had never really put together of people will immediately when they see a bigger number on something, whether you're higher cost or bigger number on bigger number of risk, it affects them differently.
00:29:20.160 --> 00:29:26.799
Uh so I've been using the scoring system a lot, and people have never had anyone come back and disagree with it.
00:29:26.960 --> 00:29:33.359
They're like, oh yeah, that's that's that's third party risk, that's uh FERPA data or HIPAA data.
00:29:33.519 --> 00:29:37.680
Oh yeah, if it gets out, that's going to be huge reputational and huge financial.
00:29:37.839 --> 00:29:40.000
Oh, yeah, 600, that's a big number.
00:29:40.079 --> 00:29:41.839
We got we got to do something about that.
00:29:42.079 --> 00:29:48.960
So it's it's a little bit um it's social engineering, social engineering leadership, but I think that that's important though.
00:29:49.039 --> 00:29:59.519
I mean, you have to know what your customers are and you have to know what your people are doing and how those pieces will affect them to help guide their decisions.
00:30:00.240 --> 00:30:01.200
Absolutely.
00:30:01.440 --> 00:30:05.839
So
let's let's talk a little bit more about some governance stuff.
00:30:06.079 --> 00:30:11.599
Um, you know, it the the the idea of governance is scary to a lot of people.
00:30:11.680 --> 00:30:16.160
Um, and they may not have a governance person or even a risk person.
00:30:16.559 --> 00:30:26.640
How would you help a CISO like me or an IT director or somebody who they don't do this every day, but they know it's important.
00:30:27.119 --> 00:30:29.200
How what what is your guidance to them?
00:30:29.359 --> 00:30:35.200
How can how can we help them know they need to do this and feel comfortable doing it?
00:30:35.519 --> 00:30:36.240
Yeah, yeah.
00:30:36.400 --> 00:30:39.039
So definitely two different halves of that question.
00:30:39.200 --> 00:30:41.200
How do we make sure that they know they need to do it?
00:30:41.359 --> 00:30:50.799
And then two, of course, how do they actually do it in a way that's realistic and accessible and you know, not overly daunting.
00:30:50.960 --> 00:30:56.400
And so I think on the first half of it, I think it's about education to make sure that they know that they can do something.
00:30:56.559 --> 00:31:02.880
I don't know a single CISO or security professional or board member who's not thinking about the risk of AI right now.
00:31:03.200 --> 00:31:07.200
I think that what they are thinking about though is like, what do I do about it?
00:31:07.440 --> 00:31:10.880
And so I think it's education to say, look, you can do something.
00:31:11.119 --> 00:31:21.759
Because I do think that the response that you mentioned earlier of just banning AI products, it's still them trying to do something, but it's not necessarily something that is sustainable.
00:31:22.079 --> 00:31:35.680
And so I think it becomes a question of saying education on look, there are there are a number of you know paths and steps to take to make sure that the organization is secure in its use of AI and that you can offer this.
00:31:35.839 --> 00:31:44.720
And you need to balance that innovation with the security that's needed, something that I'm very used to doing, as I mentioned from my previous experience.
00:31:44.799 --> 00:31:48.160
So I really understand the weighting of those types of things.
00:31:48.319 --> 00:31:49.920
But that's what that's what we need to do.
00:31:50.079 --> 00:31:51.680
And then we need to get to the how.
00:31:52.400 --> 00:32:03.519
I think that as a you know, a team that has very little GRC support internally, um, it is really scary to start on this AI governance path.
00:32:03.680 --> 00:32:15.039
And I think that you've got to partner with folks then who know what they are doing and who have the tools, whether that's a platform, an agent, or services to help you get there.
00:32:15.279 --> 00:32:17.039
Of course, we are one option.
00:32:17.200 --> 00:32:18.720
It sounds like you are one option.
00:32:18.799 --> 00:32:20.960
So it's like there's these folks out there that know.
00:32:21.119 --> 00:32:25.359
You have to ask them the right questions, make sure that they're actually going to be able to support you.
00:32:25.599 --> 00:32:33.680
Make sure that the weight's not gonna fall on you to do it, but you also need to move semi-quickly because these things do take a little time to stand up.
00:32:33.920 --> 00:32:35.279
And the time is now.
00:32:35.440 --> 00:32:40.160
Uh, and so that's kind of my my very honest advice for folks who are strapped internally.
00:32:40.319 --> 00:32:46.880
Of course, we could talk about how to make that business case, but um, that's that's I think the the core of it.
00:32:47.440 --> 00:32:51.119
Yeah, and I want to reiterate for the audience what you said there.
00:32:52.000 --> 00:32:59.279
We, our jobs, um, and even the job of a CISO or even an IT director, we're that subject matter expert.
00:32:59.440 --> 00:33:04.480
You know, we we should be the smartest person in the room about that particular topic.
00:33:04.640 --> 00:33:16.480
Um, and I try to get CIOs and IT leaders and even security leaders to understand that they're not always going to be that SME, but you can find them.
00:33:16.640 --> 00:33:21.359
Go and find that SME, trust them to come in and do it.
00:33:21.440 --> 00:33:26.799
And and I'll tell you, that's really hard for CISOs and CIOs to say, I don't know that.
00:33:26.960 --> 00:33:28.720
I don't know that thing.
00:33:29.440 --> 00:33:32.400
And what I try to get across to them is that's okay.
00:33:32.559 --> 00:33:38.799
There's nothing wrong about not knowing that thing, but there's something wrong about not asking for help.
00:33:38.960 --> 00:33:47.039
And the more you ask for help, the more you get that experience, the more they find someone like you to guide them through these steps because you do it every day.
00:33:47.200 --> 00:33:51.839
Um, no one would ever expect them to be a plumber or to be a carpenter.
00:33:51.920 --> 00:33:56.400
Their job is to be IT or IT security or whatever it is.
00:33:56.640 --> 00:34:02.559
So I think that, you know, for the audience, I would like to reiterate that there are people out there that can help.
00:34:02.640 --> 00:34:06.880
Just go and ask, you know, that there's and this is still new.
00:34:07.039 --> 00:34:11.280
I mean, I remember the first AI uh implementation engagement I did.
00:34:11.440 --> 00:34:15.280
This was, I don't know, a year and a half ago, two years ago, whenever it was.
00:34:15.519 --> 00:34:17.280
Um, this was still new.
00:34:17.440 --> 00:34:27.840
This was very, everyone was there was nothing, you know, there wasn't samples out there, there was not questionnaires out there, there was not sample vendor agreements out there.
00:34:28.000 --> 00:34:30.559
Now you can get that stuff and now you can do the research.
00:34:30.639 --> 00:34:35.199
And but it takes time and effort to really absorb that.
00:34:35.360 --> 00:34:40.400
So I think that that's that's really important for people to understand is go and ask for
help.
00:34:40.559 --> 00:34:42.800
You know, that's that's why SMEs are out there.
00:34:43.440 --> 00:34:44.800
So go Andre.
00:34:45.760 --> 00:34:46.079
Okay.
00:34:46.639 --> 00:34:52.079
Um, let's take it to the next level for that, for that CIO or that SME.
00:34:52.320 --> 00:34:54.239
They they they need that help.
00:34:55.119 --> 00:34:59.119
Where do they you you talked about that, you know, find that risk.
00:34:59.199 --> 00:35:00.239
And and I agree with you.
00:35:00.320 --> 00:35:04.719
I think that third party management is going to be that huge risk.
00:35:06.159 --> 00:35:07.920
How do I go ask questions?
00:35:08.719 --> 00:35:10.639
I'm a I'm a I'm a I'm a CIO.
00:35:10.719 --> 00:35:12.239
I I I deal with the computers.
00:35:12.400 --> 00:35:15.280
How do I ask those questions in the organization?
00:35:15.440 --> 00:35:23.039
What's what is your your hand holding, your best advice to them is like, hey, go ask these people these questions.
00:35:23.440 --> 00:35:25.199
For the third party side?
00:35:25.519 --> 00:35:26.000
Yeah.
00:35:26.239 --> 00:35:26.639
Yeah.
00:35:26.880 --> 00:35:27.119
Yeah.
00:35:27.199 --> 00:35:30.880
I mean, I'm a bit biased because I think that it needs to be automated, not manual.
00:35:31.199 --> 00:35:35.119
So I think from my from my perspective, it is really quite easy.
00:35:35.280 --> 00:35:51.199
I mean, to be honest with you, like I think that using a tool, you know, that actually has those questionnaires ready to go can send them out en masse to your vendors and basically have then an AI agent built in to review their answers, which instantly flag risks to you.
00:35:51.360 --> 00:35:52.159
That's what we do.
00:35:52.320 --> 00:36:00.239
And and that's just, I'm biased towards it because I feel like it's so efficient that I feel like, you know, it's such a good way to get this done quickly.
00:36:00.480 --> 00:36:08.639
And then in order to enable that, it's really a question of saying, okay, well, do we have a list of our vendors, which sometimes is a battle, by the way, to get that.
00:36:08.800 --> 00:36:11.599
So then I would say, okay, go to your procurement system.
00:36:11.679 --> 00:36:12.480
Hopefully you have that.
00:36:12.559 --> 00:36:17.119
If worst case, go to accounts payable and get that list and then work backwards.
00:36:17.199 --> 00:36:18.079
And that's literally it.
00:36:18.159 --> 00:36:21.920
Because all we need is the list of vendors and list of contacts.
00:36:22.079 --> 00:36:24.159
Forget even the names of the vendors.
00:36:24.320 --> 00:36:25.519
You know, we just had a list of the contacts.
00:36:25.599 --> 00:36:28.079
We could even get it, just get the assessment done that way.
00:36:28.320 --> 00:36:33.760
And within literally seven days, you would have significant more visibility than you have right now.
00:36:33.920 --> 00:36:39.599
And then you would have a second wave of visibility come in when you know all the people who needed to be chased,
um, got answers.
00:36:39.760 --> 00:36:41.760
And again, the our agents do that end to end.
00:36:41.920 --> 00:36:43.519
So that's that's actually what happens.
00:36:43.679 --> 00:36:47.920
And you end that process with the list of all your findings against every vendor.
00:36:48.079 --> 00:36:50.079
And that is just amazing.
00:36:50.239 --> 00:36:53.280
So that's the short, that's the short version to be honest with you though.
00:36:53.519 --> 00:36:58.159
Well, and the reality is IT people don't get in IT because they like working with people.
00:36:58.320 --> 00:37:03.440
So if you have a tool that can help automate that to reach out, ask those questions.
00:37:03.599 --> 00:37:05.199
I mean, I think people are about to jump on that.
00:37:05.280 --> 00:37:06.880
They they love that sort of thing.
00:37:07.119 --> 00:37:11.920
Um, the other thing is, you know, like you said, query your your systems.
00:37:12.000 --> 00:37:16.559
If you've got a contract management system, you've got an ERP that has payments.
00:37:16.719 --> 00:37:23.199
I mean, those are things, just go pull with permission and with the appropriate people involved, but pull a query.
00:37:23.440 --> 00:37:30.639
Definitely, if anything says GPT in it or AI in it, you can immediately go, okay, I need to look at that.
00:37:30.800 --> 00:37:33.199
So yeah, no, that's that's good.
00:37:33.360 --> 00:37:43.679
I, you know, one of the things that when I'm I'm dealing with the AI with the with IT leadership, they uh number one, they don't want to do to compliance.
00:37:43.760 --> 00:37:45.360
They it it just scares them.
00:37:45.440 --> 00:37:49.199
Um, and they don't like really talking to other business units.
00:37:49.599 --> 00:37:52.400
Uh, and it's really important for them to do that.
00:37:52.480 --> 00:37:58.480
So I've got this whole series I'm doing this year for CISOs and IT leadership for that social side.
00:37:58.719 --> 00:38:00.719
CISO is a diplomat, is what I call it.
00:38:00.800 --> 00:38:03.039
And it's just that how do you engage?
00:38:03.119 --> 00:38:07.519
And um, and I did my presentation the very the very first time.
00:38:08.079 --> 00:38:13.840
Uh I got to the middle part of us, all right, everyone, this is gonna feel really creepy.
00:38:14.719 --> 00:38:16.239
But you're gonna have to do it.
00:38:16.400 --> 00:38:23.280
I want you to sit down and I want you to think about all the people that you need to to talk to.
00:38:23.440 --> 00:38:31.280
And I want you to write their names down, and I want you to make it an intent that you will talk to them this week, even if it's just to say hi or whatever, whatever it is.
00:38:31.360 --> 00:38:36.239
So it's just this guiding them to go and ask questions and to engage and building influence.
00:38:36.400 --> 00:38:39.599
And I had several people come up and it's like, yeah, we don't like doing that.
00:38:39.679 --> 00:38:40.559
We we don't want to do that.
00:38:40.639 --> 00:38:43.920
It's like,
you're not gonna be very successful at this job.
00:38:44.000 --> 00:38:49.039
You really need to, so anyway, I'll I'll I can get off on that all day long.
00:38:49.199 --> 00:38:57.039
But I do have another question for you before we kind of get to uh uh towards the end of this is you you mentioned the e the EU AI Act.
00:38:57.199 --> 00:39:02.719
And there are some laws that are coming out about AI in different countries and different things.
00:39:02.960 --> 00:39:12.800
When when GDPR came out and data privacy really started to become a thing, GDPR was kind of the that that big boy, that one that got everyone's attention.
00:39:13.039 --> 00:39:16.079
The EU AI Act is kind of the same way.
00:39:16.239 --> 00:39:18.880
People are really standing up and taking attention to that.
00:39:19.119 --> 00:39:37.760
What are what do you believe, and based on your experience, is going to happen here in the US for AI either at a state level or a federal level or even a sectoral level where these different industries are going to require differ you know compliance, you know, laws, rules, guidance, whatever it may be.
00:39:37.840 --> 00:39:40.239
What what are you seeing might happen?
00:39:40.960 --> 00:39:46.400
Well, thanks, first of all, for bringing up the privacy side by side, because I think that's exactly the right analogy.
00:39:46.880 --> 00:40:01.760
I would I would also push a little bit into looking at um just kind of like SOC 2 and ISO 27001 as another analogy, kind of baseline security standards uh as well as privacy and then comparing it to AI.
00:40:02.079 --> 00:40:17.360
I think that the EU AI Act is going to be a uh very similar force as GDPR, in that I think we'll see, you know, a lot of different AI privacy acts aligning to the principles of that as kind of the first mover.
00:40:17.519 --> 00:40:22.400
I think the EU AI Act is already a bit stale, which is tough, but I respect that it came out fast.
00:40:22.880 --> 00:40:31.760
I think that the US is going to see more state-by-state regulation, which sucks for GRC teams in the US or anybody serving the US market.
00:40:31.920 --> 00:40:35.199
Uh, but that is the way of privacy regulation as well in the US.
00:40:35.280 --> 00:40:36.480
So that just is what it is.
00:40:36.639 --> 00:40:37.119
Yes.
00:40:37.760 --> 00:40:43.199
We have, for example, a state privacy tracker that cross maps all of the state privacy requirements.
00:40:43.280 --> 00:40:54.480
And we're doing the same thing for AI right now to be able to say, look, here's one set of requirements that covers EU AI Act, you know, and other state um level AI privacy regulations.
00:40:54.559 --> 00:40:55.519
So that's useful.
00:40:55.760 --> 00:41:11.119
The last thing I'll say though is that with kind of general security standards, so SOC 2, ISO 27001, I find it refreshing and kind of great that the market has driven those standards to be required rather than regulation.
00:41:11.440 --> 00:41:30.239
And I think that the US and the kind of nature of where things are right now, people are so aware of this AI risk that I do think that they are starting to self um, you know, self-search for providers and so on that are being more careful.
00:41:30.400 --> 00:41:37.280
And I think that that's the first steps towards the level of kind of uh requirement that SOC 2 has seen, for example.
00:41:37.440 --> 00:41:38.719
So that's my very honest answer.
00:41:39.679 --> 00:41:40.480
No, I like that.
00:41:40.639 --> 00:41:41.360
I agree with that.
00:41:41.519 --> 00:41:57.920
I think it could be difficult to get, you know, big laws, you know, here in the United States, but states, you know, it's just like Apple GDPR, you know, California came out with uh with a variety of laws, the CPA, the CCPA, the the CRPA, things like that.
00:41:58.159 --> 00:41:59.599
I think it's gonna be the same.
00:41:59.760 --> 00:42:08.320
I think there's gonna be some restrictions that come in for AI and they're gonna pass it, either unfortunately through fear or otherwise, you know, bad things happening.
00:42:08.480 --> 00:42:11.280
But I think that that's probably gonna it's gonna happen.
00:42:11.360 --> 00:42:12.480
So I agree with you.
00:42:12.719 --> 00:42:18.079
So we've only got just a few minutes left, and I really want to hear more a little bit about your your company.
00:42:18.159 --> 00:42:19.440
You know, but what do you do?
00:42:19.599 --> 00:42:24.960
How, what is it that you provide that the audience here can really take advantage of?
00:42:25.119 --> 00:42:29.519
Because I'm sure they've listened to this and they're like, I don't even know where to get started.
00:42:29.679 --> 00:42:34.079
You know, I they've said a few things and I agree with them, but I just don't know what to do.
00:42:34.159 --> 00:42:36.559
So what what's your what's your elevator pitch?
00:42:36.880 --> 00:42:37.840
Yeah, absolutely.
00:42:38.079 --> 00:42:42.880
So at the end of the day, the compliance platform is a full suite GRC product.
00:42:43.039 --> 00:42:48.800
We support controls management, risk management, third-party risk policies, and customer trust.
00:42:49.039 --> 00:42:59.920
The difference, though, with our platform and the other GRC products out there is that we use agents for about three years now to actually do the work.
00:43:00.159 --> 00:43:03.920
So most softwares, you buy the software and you do your work in the software.
00:43:04.079 --> 00:43:09.760
Our software does the work itself and it frees you up to do the work that you didn't have time for before.
00:43:09.920 --> 00:43:13.599
And I really like to talk in GRC about the important versus the urgent.
00:43:13.840 --> 00:43:17.519
Everything in GRC is so urgent, so urgent all the time.
00:43:17.760 --> 00:43:21.519
But when your stack of urgent things is so big, you never get to the important.
00:43:22.079 --> 00:43:29.760
And the important things include supporting the building of a security culture, include things like educating yourself on AI governance.
00:43:30.000 --> 00:43:36.320
And so we actually take the work off of GRC teams' plates with our agents, proved and tested.
00:43:36.400 --> 00:43:41.519
Our clients include CVS Health and Major League Soccer, Dropbox, so many more.
00:43:41.679 --> 00:43:43.679
And they use our agents every day to do work.
00:43:43.760 --> 00:43:46.480
And so that's a little bit about how we're different.
00:43:46.719 --> 00:43:52.159
And we, you know, use that for things like AI governance, all that third-party risk, it's so much more.
00:43:53.360 --> 00:43:55.039
Well, I think that's fantastic.
00:43:55.199 --> 00:44:01.119
And again, we kind of go back to what we said before is sometimes you got to talk to that SME and help them guide you.
00:44:01.199 --> 00:44:10.480
And I think that sounds like, you know, the platform that you guys have, the tools that you uh have can really help people get started and/or progress.
00:44:10.639 --> 00:44:21.360
You know, it may they may already have a pretty good GRC program, but doesn't mean that people could be a little bit more efficient and and have more things standardized in their in their processes.
00:44:21.519 --> 00:44:22.880
So well, that's fantastic.
00:44:23.039 --> 00:44:24.559
Well, Richard, thank you so much.
00:44:24.719 --> 00:44:29.519
I know you're traveling and you've got to run soon, but thank you again for jumping on the podcast.
00:44:29.679 --> 00:44:31.679
Uh, I really enjoy this conversation.
00:44:31.840 --> 00:44:36.239
I love talking with other privacy professionals and other AI governance professionals.
00:44:36.320 --> 00:44:49.119
It's just a it's refreshing to have that person who uh you can talk to that understands the stuff, the way we do, and the the desire we want to have for the community and the industries to get better.
00:44:49.280 --> 00:44:50.960
So thank you so much for coming on.
00:44:51.199 --> 00:44:52.159
Oh, it's my pleasure.
00:44:52.320 --> 00:44:53.440
Thanks, Jonathan.
00:44:53.840 --> 00:44:54.960
Absolutely.
00:44:55.280 --> 00:45:02.239
With that, uh, we will go ahead and end the Secure AF podcast, and we're uh excited for the next one to come very soon.
00:45:02.400 --> 00:45:05.840
So you guys have a great rest of the week, and we'll see you next time.
00:45:06.000 --> 00:45:07.039
Thanks, everybody.
00:45:07.280 --> 00:45:12.000
The Secure AF Podcast is a production of Alias Cybersecurity.
00:45:12.159 --> 00:45:15.920
Visit us online at aliascybersecurity.com.
00:45:16.079 --> 00:45:17.760
All rights reserved.