How AI is reshaping cybersecurity with Cisco CEO Jeetu Patel

So let's talk about AI and cyber attacks. So artificial intelligence is changing how we work. But what happens when hackers use it, too? With the midterm elections approaching, we're looking at the risks and how to protect yourself. Cisco President G2 Patel joins us live. And G2, I really appreciate you being here this afternoon. Thank you for having me, Tyler. How are you doing? I'm doing pretty good. And a lot of us use AI, whether consistently throughout the day or at least once or twice throughout the day. I feel like a lot of people have really embraced this technology, but obviously AI can be dangerous in the wrong hands, especially with hackers. And so how quickly are they able to utilize this type of technology to be able to take advantage of people? Pretty quickly. So if you take a step back and say, what is happening? Most people are where you just described right now, which is we've been using AI from the form of like chatbots intelligently answering our questions. You ask it a question, you get back an answer. It actually has been been pretty great. We are now moving to that second phase, which is these personal agents and agents that actually conduct tasks and jobs for us almost fully autonomously. And they might work in the background, 7 by 24, Tyler. And so what you have is this exposure of having cyber attacks gets to be far greater because these agents are working on a continuous basis in the background. And you're going to give it much, much more access to information. You might give it access to your email. you might give it access to your calendar, you might give it access to your credit card because it'll shop on your behalf, whatever you might need it to do. And so that gives you, it's important that consumers take a moment and say, how do I protect myself so that we don't have unnecessary levels of risk that can get exposed to it? And you have to mitigate that risk. And so are there some specific weak spots in terms of AI that people can target or that hackers can target yes i think what there's two aspects of it the first one is when you what you said which is bad actors or adversaries or hackers who will actually try to go out and corrupt a model uh and jailbreak the model so that the agents can themselves uh have behavior that you was unintended right and that might be bad for people um that's one dimension but the second dimension is when you have these agents that are they're kind of like teenagers tyler so think of them they're supremely intelligent um you know and they have no fear of consequence and frankly they exercise poor judgment far more frequently than you and i would like and so what you want to do and that second part is when you that agent might just feel like hey you asked me to do something i'm just going out and doing it but i haven't really applied common sense. And so what you have to do is make sure that you distinguish between those two. And these have to be dealt with separately. So on the second one, giving an agent your credit card and asking your agent to shop on your behalf, but not giving it guardrails where the agent can just start to go nuts and say, hey, you asked me to go get a sweatshirt. I'm going to get you a $500 sweatshirt because you never told me any boundary conditions. Those are things that you have to also be careful of. So I feel like there's there's two different dimensions going on. One is hackers intentionally corrupting agents. And the second is, you know, these agents being extremely, you know, literal in the in the way that they take orders from you. And so I'd have to talk about this. We were talking about the election a little bit ago. J.D. Vance, Vice President J.D. Vance is here in the Atlanta area right now with the elections approaching. How quickly can AI be used and how can it be used to be able to honestly mislead voters? And does that concern you? I think what you have to keep in mind is there's going to be a fair amount of creativity that's applied to causing harm because all these hackers and adversaries have the same tools that they have access to that you and I have to be productive. So one of the things to keep in mind is these attacks are going to get much more machine scale. And when they get machine scale, which means the scale proportion of these attacks are going to be at very, very fast speeds and many, many more of them than what we've been used to seeing in the past, what you have to have is your defenses also have to get machine scale. So here's the counterintuitive portion. People think AI is actually going to make you unsafe. But if you don't use AI in cyber defense, you will be the most unsafe you've ever been. Because if your attacks are happening with AI and you actually use humans to go out and defend against it, AI is just going to beat you out. And so you have to use AI for the cyber defense side, you know, kind of side of things. And so what are three things people who are listening right now, watching and joining us, what are three things that viewers can do today to be able to protect their accounts and personal information when it comes to AI and obviously security? The first thing you have to do is you have to get familiar with AI. Sitting on the sidelines not knowing what's going on is not a good idea. So you have to get familiar with it because there's going to be huge implications to society with AI. So that's my first piece of advice. Don't sit on the sidelines. But two, as you start using these technologies, be very careful of the kind of access you grant these technologies. Because oftentimes, depending on the products you're using, you want to make sure that you're going into the settings and ensuring that you're not over-provisioning permissions to these systems. Don't give it your credit card to start with. Don't give it your bank account number to start with. Have very kind of tight controls on it, and make sure that when you're providing instructions to your agent to say, go do something on my behalf, that you're giving it very precise instructions and not having it go outside of the bounds. So that's the second thing. The third and most obvious thing, Tyler, that we've actually seen happen quite frequently is people don't use two-factor authentication. So they just have a user ID and a password, and password is very easy to get hacked. So something very simple is just use a second factor of authentication so that you have a code that is asked after you put your password in that you have to put in so that the computer knows that it's you, not someone else doing it. Well, G2 Patel, Cisco president, I appreciate you joining us. this afternoon to discuss AI and obviously the power of AI and why people need to be obviously paying attention to what exactly is going on with it and what's how people can use it for bad. So I appreciate your time breaking all of this down for us. Thank you for having me, Tyler. Take care.