VicOne’s Post

Transcript

Let's go through some examples and I actually have two examples over here. So the first one is taking that previous picture about marrying this intelligence that you have and then with automotive threat intelligence that maybe it's kind of new to you, right? So let's go through this, you guys. So here we're launching the LM solution and maybe you're looking at through intelligence as you start your day. And so the solution that you would have to kind of go through would kind of speak about maybe open source threat intelligence or maybe the dark web information. And let's say you're going through information about recent research pertaining to a certain vulnerability. The first thing you're going to ask is. Well, what is the scope of this, right? So you're going through what firmware and are there any things that have happened in the past that's actually related to this newer threat intelligence? After that mitigation measures are this mitigation activities that were kind of spoken about. Maybe you're thinking about information about the dark web. If you want to protect your users, maybe you have that kind of information or concern already, then you can actually have this. Um, this last question in blue, you could see that it's like a, an off kind of query, but the reason behind there is it's one of those things that any one of you here that's looking to generate or actually produce such technology, you would actually have to have safeguards around that question, right? So it's very important to have that contextual kind of information centered on intelligence. And so if. These prompts are these questions are not relevant to what you're looking at and make sure that the solution that you're working on or you're trying to build a response appropriately. So I have this screen shot up and this is again, the blue colored boxes are the prompts and the Gray colored boxes are the responses from a generative AI solution. And this is just an example, right? If you're going through sort intelligence. As a human, one of the things you're going to have to do is to go through a lot of this data, read, identify, maybe, maybe take notes, watch a video or whatnot. And if you could get the summary and how it actually affects you, that is essentially your end goal. Right. So if you're going through threat intelligence, if you're going through a newly newly published CVE, how does it affect you? And that's entirely a question. And what is the scope of that, right? So the relevance or the this contextual information is quite important so that you don't have to spend what, 3045 minutes going through things, looking at an inventory list, if you can have the series of prompts that would actually bring you this information. It would be much faster for you to. To go about and maybe update your stakeholders about this newer vulnerability that actually out there. So the other thing also that would be interesting and this is something that you may want to consider because it's actually out outside of any vehicle or data. Um, increasingly in the software defined vehicles, you know, the, the users are now being tied to the vehicles itself. So they would actually have user username and password prompts and said mobile app or anything that would have to identify the user type back into the vehicle. And with this identity, that's often one of those things that red actors are interested in because. If a user name password is actually compromised, then you could think of very creative queries and how to actually utilize that, and these are even outside of the automotive industry. Identity is one of those frequently shopped aspects that are being sold in the dark web. And so if you have methods of acquiring such data and you could plug it into your generative AI solution, it would be very good if you could make a play for that, Make a use case for that so that you can now think of further ways of protecting your end users. So this is one of those things that you may wanna investigate. Um, so I mentioned a while back that the generative AI actually lives and breathes and data and operate more efficiently when they have a lot of this data ingested because they could make sense of it, right? And they could contextualize their response to the end user. The the idea that you may want to go for. Is this concept of a single pane of glass? And I know this is an overly used term, but still it doesn't mean to say it's wrong, right? The idea of this single pane of glass is that it's a centralized visibility for virus sources of data. Stevie's or maybe you know, vehicles are entering this era would generate a ton of information that can be used often terminals, logs, data telemetry if you will. And they would generate data for each and every activity in the vehicle. But not only that, there's also a ton of data in the back end infrastructure that is supporting said vehicle or this fleet of vehicles and so. Once you have the data, you'd have to think about, OK, what can I do with it? Well, the first thing that first thing that comes into mind is actually do this detection response phrase. And you would have to respect that the vehicle actually has this light cycle, the product life cycle that goes through design, development, test production and on the road that goes about for like 7 or 10 years, right? And each of these spaces would generate tons of data which which. And then support a lot of this. So you have data that's already been generated by the vehicle and maybe data to you're able to collect through the entire life cycle. You would then have to think what are these technologies that you may already have in place? Maybe it's not the technology, maybe it's a process, right? Think of it as well, can you link one process to the next in a seamless manner? Another way to term this is like orchestration and automation. Automotive threat intelligence when those fundamental aspects of this. This approach, so you'd have to have automotive threat intelligence then if you put in this information about or this implementation of generative AI, you can then take all of this foundational technologies throughout the entire life cycle, look into the existing data that's being generated by the vehicle then think about performing some sort of risk management, automotive risk management so. Umm. We respect that. You have to start from somewhere, you know, maybe you don't have access to a certain part of data. That's great. I don't think anyone's gonna get this right the first time, to be honest. But it is a picture that you could paint and kind of work through so that you could have this. The single vision of where you want to go.