1
00:00:00,000 --> 00:00:09,000
Welcome to the show everybody. Today is Wednesday, July 24th. We're broadcasting here live from Research Triangle Park in North Carolina.

2
00:00:09,000 --> 00:00:16,000
Got Sue Deer here as one of the guests. He's actually sitting right beside me. You can't see him on my screen, but you can on his.

3
00:00:16,000 --> 00:00:22,000
Andres, good to see you as always. We have a fantastic show today on artificial intelligence.

4
00:00:22,000 --> 00:00:28,000
Not just on AI, but specifically the security aspect of artificial intelligence.

5
00:00:28,000 --> 00:00:36,000
Everybody claims to have AI now, and we see it in every advertisement related to technology. It's on our phones. It's in our cars.

6
00:00:36,000 --> 00:00:40,000
It's very cool and very exciting, and it's going to remain that way.

7
00:00:40,000 --> 00:00:50,000
It's going to remain that way until it gets majorly hacked and we're all scrambling. So, as we security-minded people know, security is usually an afterthought

8
00:00:50,000 --> 00:00:58,000
when new exciting technologies get rushed to the market, and AI is absolutely not an exception for that.

9
00:00:58,000 --> 00:01:07,000
Exactly. Usually and unfortunately, it's always the last thing that we think about. And when we're creating new technology,

10
00:01:07,000 --> 00:01:13,000
look at the examples that we've seen from cloud technologies early on.

11
00:01:13,000 --> 00:01:22,000
How to secure those, all the hacks that happened when it was maturing and it is, I still say that it's still maturing.

12
00:01:22,000 --> 00:01:29,000
So, it is an interesting thing to think about security for a new technology.

13
00:01:29,000 --> 00:01:40,000
Now, companies are already using AI. Most of them have already thought about how to use AI, but a lot of questions come up.

14
00:01:40,000 --> 00:01:49,000
Are they exposing any type of information? Is it secure? Is it safe? And that's basically what we're here to do today.

15
00:01:49,000 --> 00:01:57,000
We're going to talk about some of those things. I'm excited that we have Joel Sprague and Sudhir Desai on the show today.

16
00:01:57,000 --> 00:02:08,000
These guys know a lot of what's going on with AI, and we're going to talk about these concepts, and it's going to be a super, super, super nice conversation today.

17
00:02:08,000 --> 00:02:16,000
So, with that, I'll pass it to you, Joel and Sudhir, to introduce yourself and get started with the show.

18
00:02:16,000 --> 00:02:25,000
Sounds good. Thanks, Andres and Mike. I am Joel Sprague. I am an account SE here at Cisco. I've been here about five years.

19
00:02:25,000 --> 00:02:37,000
Prior to coming here, I ran a private cloud for a Fortune 100 and, you know, have basically been in IT for 25 years and continue to, you know, enjoy the fact that there's always something new to learn.

20
00:02:37,000 --> 00:02:54,000
And how to secure it. Sudhir? Indeed. My name is Sudhir Desai. I am also an SE, however, comma security. So, security generalist SE, I guess, whatever our names are this week.

21
00:02:54,000 --> 00:03:07,000
But I've been in the sales role for about five years now, 2018, however long that was. I'm not really counting 2020, so I think that's like five years or so.

22
00:03:07,000 --> 00:03:13,000
But before that, I was over on the support side, came in through the Sourcefire acquisition.

23
00:03:13,000 --> 00:03:22,000
So, I have seen the fall rise, fall rise of the firepower work set. So, very nice to meet you guys.

24
00:03:22,000 --> 00:03:38,000
All right. Well, we're super excited to have such talent on the show. All the brain power. Andres, I'm interested, you know, if we combine Sudhir and Joel and we compare them to like an artificial engine, like, I don't know who would win that battle when it comes to the brain power there.

25
00:03:38,000 --> 00:03:44,000
So, we're going to find out today, though. A lot of cycles on that one.

26
00:03:44,000 --> 00:03:56,000
One of the questions I would like to throw out there just before we get started into the security aspect, Joel, I'll throw this to you to start it off. Just in general, how does AI work? What are the components?

27
00:03:56,000 --> 00:04:01,000
Maybe talk a little bit about where the data actually lives.

28
00:04:01,000 --> 00:04:17,000
Sure. So, you know, AI is a huge field that encompasses a lot of things, but here in 2024, that's a fancy keyboard. Okay. Here in 2024, what we're generally talking about is generative AI.

29
00:04:17,000 --> 00:04:31,000
The kind of the idea behind generative AI is that rather than just predicting things, it is creating new material. You know, this is your chat GPT, your llama, everything like that.

30
00:04:31,000 --> 00:05:00,000
So, I like to describe generative AI as really fancy, expensive, autocomplete. The way we build these models is we take a huge corpus of data. So, if you're talking about a medical AI, you might take every published study in PubMed that, you know, has hit all the medical articles, and you feed it through an algorithm that what it's trying to do is basically predict what the next word is going to be.

31
00:05:00,000 --> 00:05:18,000
It's autocomplete. So, this data is fed through the AI and is then compared to a training set, or no, that's the training sets compared to the test data set where it's we know what those values should be.

32
00:05:18,000 --> 00:05:42,000
And the AI says, okay, I got it about 70% right, let's change this, this and this in the algorithm and run it all through again. And that's really kind of, you know, what makes AI so interesting is that it's not like typical software where humans are writing the programs, we're just saying, here is all this data, we want to classify it.

33
00:05:42,000 --> 00:06:00,000
You know, with LLMs, it's what's the next word, you know, basically, go do what you're going to do, you know, change your algorithms, do whatever you have to do until your prediction of what the next word will be matches up to this set where we know the correct answer.

34
00:06:00,000 --> 00:06:17,000
We don't care how you get there, we know that it's, you know, a huge multi vector database that a human is never going to understand. That's fine. Just train yourself to get to the point where you can match the predictions from these known values in this test set.

35
00:06:17,000 --> 00:06:24,000
And that takes a whole lot of GPUs and a whole lot of electricity and cooling to get there.

36
00:06:24,000 --> 00:06:28,000
We're not running this on like a Raspberry Pi then.

37
00:06:28,000 --> 00:06:44,000
We're running these large language models. The fun thing is, though, once we do train these models, and there are small language models and other things as we're starting to realize that, you know, not everyone can buy a data center co located at a nuclear reactor like Amazon to power it.

38
00:06:44,000 --> 00:06:59,000
We are coming up with smaller models but even then, you still need the GPUs and this massively parallel computing first to train these small model. But then yes, we know we're starting to get ones we can run on smaller things like even running them on some of our catalysts

39
00:06:59,000 --> 00:07:04,000
9300 switches.

40
00:07:04,000 --> 00:07:08,000
That's pretty cool. I didn't know that.

41
00:07:08,000 --> 00:07:19,000
The joys of the fact that we can, you know, basically just have Docker on there and whatever random idea someone gets it's let's build a Docker image and push it to a catalyst.

42
00:07:19,000 --> 00:07:24,000
There's an AI for that.

43
00:07:24,000 --> 00:07:27,000
That was really good man. Thank you. Thank you so much.

44
00:07:27,000 --> 00:07:35,000
I guess for for the next question, this one's going to be, you know, something that everybody just has in mind.

45
00:07:35,000 --> 00:07:39,000
We, we keep saying and I know we keep saying it to our customers.

46
00:07:39,000 --> 00:07:51,000
AI it's been here, and it's been here for a long time but why are we just seeing it right now like what's, what's the deal and so there if you don't mind going over a little bit about that.

47
00:07:51,000 --> 00:07:55,000
Why do we see that now.

48
00:07:55,000 --> 00:08:02,000
It's kind of funny to me that there's so much more emphasis these days put on AI.

49
00:08:02,000 --> 00:08:07,000
Because we just are more inventive with our buzzwords these days.

50
00:08:07,000 --> 00:08:22,000
I say that because we've had, as Joel was mentioning, we've had AI for years, decades, even back in the day I know back when I was in college in the early the pleocene era, or wherever.

51
00:08:22,000 --> 00:08:24,000
I kid I'm not that old.

52
00:08:24,000 --> 00:08:36,000
But in any case, you know, we had list and that was doing natural language processing where we could write programs that would understand commands. We gave it such as.

53
00:08:36,000 --> 00:08:40,000
Make me a sandwich and, you know.

54
00:08:40,000 --> 00:08:44,000
If you had the processes on the physical side to make you a sandwich.

55
00:08:44,000 --> 00:08:54,000
You could actually have a computer program stick the slices of bread together, cut it, whatever, you know, like the Teenage Mutant Ninja Turtles van that does everything, including makes pizza.

56
00:08:54,000 --> 00:09:01,000
You know, you had that sort of thing already existing these days with the generative piece.

57
00:09:01,000 --> 00:09:06,000
It has become a lot easier to do just because we have.

58
00:09:06,000 --> 00:09:17,000
You know, the money more money is there because people see, oh, we can do this. We can make lots of money out of it. So they put money into it. They put the GPUs into it. They put the TPUs into it.

59
00:09:17,000 --> 00:09:22,000
And videos like 6000 core processing power chip.

60
00:09:22,000 --> 00:09:33,000
You know, one you rack, but they say, cool, you know, you can heat up the sun, but you'll be able to do all this work with it. And we do all the work with it. And so we have the capability.

61
00:09:33,000 --> 00:09:38,000
To really very, very.

62
00:09:38,000 --> 00:09:46,000
Fastly, it's not a word, but we accelerate the programming of the AI and what I can do and.

63
00:09:46,000 --> 00:09:54,000
To that same end, the algorithms being used, we may still be using the same algorithms, but now that we have so much more processing power.

64
00:09:54,000 --> 00:10:03,000
We can use those algorithms. They don't even need to be efficient anymore because we can just throw more ram. We can throw more processors at the issue.

65
00:10:03,000 --> 00:10:09,000
So it's like, I think there was a, you know, one of the crazy super computers from back in the day has like.

66
00:10:09,000 --> 00:10:13,000
Less processing power than my cell phone.

67
00:10:13,000 --> 00:10:18,000
So it's, you know, just the economies of scale and what we can do.

68
00:10:18,000 --> 00:10:24,000
And then also the awareness side, we're getting lots more. It's like.

69
00:10:24,000 --> 00:10:28,000
Crowd strikes issue cross strike also had an issue with Linux a few months ago.

70
00:10:28,000 --> 00:10:32,000
But who uses Linux to be impacted by the issue.

71
00:10:32,000 --> 00:10:40,000
You know, I'd love to say tons of people, because I use Linux, but in reality, you know, that's 2% of the market. Maybe.

72
00:10:40,000 --> 00:10:45,000
So, you don't have any major systems that are public facing.

73
00:10:45,000 --> 00:10:52,000
So, you know, we're getting into problems, even Colonel panics where on the Windows side. Yeah, you know, something shuts down a Colonel.

74
00:10:52,000 --> 00:10:55,000
Boom, global blackout for travel.

75
00:10:55,000 --> 00:11:00,000
So it's that sort of scale as well for the AI side.

76
00:11:00,000 --> 00:11:05,000
There's so much more. It's running everywhere as we're going to talk about later.

77
00:11:05,000 --> 00:11:10,000
And so you just have more.

78
00:11:10,000 --> 00:11:19,000
And to come back to the question, the reason why it's so powerful and we're seeing it being more useful is because.

79
00:11:19,000 --> 00:11:26,000
We're actively adding more to it and feeding into that usefulness of it and feeding into that power of it.

80
00:11:26,000 --> 00:11:30,000
So it's been here, but, you know, yeah.

81
00:11:30,000 --> 00:11:39,000
Sorry for the huge answer. No, no, and that was really good. I know. I know, like, for for exactly what you mentioned that, you know.

82
00:11:39,000 --> 00:11:45,000
Just the example in crowds right now, the event that happened last week, but.

83
00:11:45,000 --> 00:11:54,000
With that example of the Linux computers and the things that happened to them on those and nobody even bad and I or, you know, notice.

84
00:11:54,000 --> 00:12:02,000
So that that that talks a little bit about the scale of what we can see and what we're seeing today. So that's pretty cool.

85
00:12:02,000 --> 00:12:13,000
So you guys have seen the Terminator movies, right? With like, the sky net and all that. This is where my mind is going.

86
00:12:13,000 --> 00:12:16,000
The dangers of AI. So.

87
00:12:16,000 --> 00:12:25,000
From Sudir's answer, it sounds like the, the, you know, AI has been around for a while. We're just now kind of hearing more about it because we're.

88
00:12:25,000 --> 00:12:33,000
What more attention on it and of course, like, you know, unlimited memory and RAM and systems are just exploding.

89
00:12:33,000 --> 00:12:40,000
Our real is sky net Joel. And if you could maybe touch on the dangers of AI a little bit here.

90
00:12:40,000 --> 00:12:50,000
Sky that's not here yet, but there's, you know, it's fine. We finally got into the point where, hey, this isn't just.

91
00:12:50,000 --> 00:12:58,000
You know, my favorite, you know, Philip K. Dick novel or something. This is starting to feel a little more real.

92
00:12:58,000 --> 00:13:08,000
So the dangers of AI, you know, AI is software. So there's always the regular cybersecurity, technical concerns, all of those.

93
00:13:08,000 --> 00:13:13,000
What AI really brings though is a lot of more.

94
00:13:13,000 --> 00:13:20,000
Moral or social concerns, I think are where a lot of the big dangers are.

95
00:13:20,000 --> 00:13:27,000
You know, I T folks don't tend to be looking at the social and or technology in general.

96
00:13:27,000 --> 00:13:38,000
You know, inventors create something to solve a problem. It's what engineers do, but rarely take that step back to say, how is this going to affect society?

97
00:13:38,000 --> 00:13:51,000
And I'm sure, yeah, I didn't watch it. We saw Barbie instead. I'm sure Oppenheimer went into this because I know Robert Oppenheimer talked about that a lot, as did Edward Teller and everyone else involved in that.

98
00:13:51,000 --> 00:13:54,000
Like all technologies, AI.

99
00:13:54,000 --> 00:14:12,000
Tends to amplify the traits of humanity that created it, good and bad. We saw this. We've seen this multiple times now with various chat bots on social media platforms where they ended up, you know, basically, as Sudi Alex called basically the Hitler bot.

100
00:14:12,000 --> 00:14:24,000
You know, they take in the bias and discrimination of the humans they are interacting with and amplify it to the point where someone has to pull the plug.

101
00:14:24,000 --> 00:14:34,000
We've seen it again and again and you know there are safety systems to try and avoid that, but it's it's something that seems to keep happening that we're going to have to reckon with.

102
00:14:34,000 --> 00:14:52,000
There's also the privacy concerns and you know the idea of intellectual property. You know, when you train something, an art generation bot on a full corpus of, you know, all the art in all these museums in the world, well,

103
00:14:52,000 --> 00:15:04,000
do the original artists get credit for the art that the bot then creates that is, you know, sure looks like something that Bob Ross or someone would have created, you know, it's tough.

104
00:15:04,000 --> 00:15:11,000
You know, there's also, of course, always the worry the robots are going to take our jobs. Yeah, it's.

105
00:15:11,000 --> 00:15:23,000
It happens. It's going to happen. It has happened. I will say this is AI has been one of the first technologies where lower skilled workers in the field are actually being helped by it.

106
00:15:23,000 --> 00:15:45,000
So we look at automated cashiers at fast food restaurants. That is technology taking someone's job, you know, the lower skilled they can take the job, but actually AI co pilots and helpers and call centers have been proven that they don't do a lot to help you're really experienced customer service reps.

107
00:15:45,000 --> 00:16:02,000
They already know how to do the job, but AI does a really good job of bringing up those tier one customer service reps to be more of a tier two or tier three, which, you know, you then have the risk of, well, if the AI is good enough, can we just get rid of them.

108
00:16:02,000 --> 00:16:15,000
There's still the job displacement, but at the current lifecycle of AI, we're at a point where for once technology is helping people of the less skilled in their particular field.

109
00:16:15,000 --> 00:16:22,000
We'll see if that continues or if, you know, the powers of B say, well, why do we still need to pay them.

110
00:16:22,000 --> 00:16:44,000
And then there's just ethical and moral issues, you know, are we giving Skynet control of weapons, you know, can drones be given autonomous ability to use weapons or do they still have to have human control on the hand before any decisions can be made, stuff like that's technologically

111
00:16:44,000 --> 00:16:58,000
possible already we as a society need to make some serious decisions as to how much we entrust to AI and what decisions we let them make without any human interaction.

112
00:16:58,000 --> 00:17:11,000
Good good point. And then I guess that brings into the oversight, like, who's going to be responsible. If the AI if an AI drone makes a decision on a on a battlefield to take some type of action.

113
00:17:11,000 --> 00:17:15,000
And it was the wrong thing to do. Like, who's ultimately at fault for that? Is it.

114
00:17:15,000 --> 00:17:19,000
The drone or the, you know, the.

115
00:17:19,000 --> 00:17:27,000
Interesting the manufacturer, the captain unit, the country. Yeah, it's stuff.

116
00:17:27,000 --> 00:17:38,000
My question is always going to be, where do we find our Sarah Connor to take this thing down after the fact. Right. And to what year we send her, you know.

117
00:17:38,000 --> 00:17:43,000
So true. All right.

118
00:17:43,000 --> 00:17:49,000
No, that was good. Now, the next question is, as for you, so there.

119
00:17:49,000 --> 00:18:07,000
We've seen a few things happening with AI as far as security goes and have we seen AI get hacked yet? Like, have we seen I know we talked a little bit about those things, but do you have examples things that you've seen happened with AI?

120
00:18:07,000 --> 00:18:19,000
Yeah, a couple of things. I'll just go through. I know we had been discussing a few of the issues with AI systems. And now, like a couple of the attacks that could happen on AI.

121
00:18:19,000 --> 00:18:39,000
One of them, they're called adversarial attacks where hackers can use examples of altered correct images to almost. Well, I want to say poison, but there's also a type of attack that's also poisoning.

122
00:18:39,000 --> 00:18:57,000
But, you know, to alter what the AI recognizes as a correct image. So maybe a cup for a cell phone and not something that drastic, but even slightly so that you have people with four arms being generated or different things like that.

123
00:18:57,000 --> 00:19:10,000
Which, you know, that could happen with all of these attacks at the end of the day, but one of the attacks is kind of using altered images to deceive those image recognition models.

124
00:19:10,000 --> 00:19:28,000
Another attack I had mentioned earlier is a data poisoning attack. And so to use AI, you have to go in and to Joel's point and I believe it was Andres' point earlier.

125
00:19:28,000 --> 00:19:48,000
Because we're using, we're trying to use generative AI, making new things up with the existing data. So with that large language model that we have on the back end, data poisoning means inserting malicious data, inserting incorrect data to then, you know,

126
00:19:48,000 --> 00:20:05,000
create a new age. Well, you already said Hitler, Joel, so I'll pick somebody else from history. I don't know, Kublai Khan, the Mughal dynasty. There we go. We'll pick another invader.

127
00:20:05,000 --> 00:20:22,000
But, you know, using data to and using malicious data to form an AI's opinion for it because it's trying to generate from its data set. So its opinion is generated by its data set.

128
00:20:22,000 --> 00:20:34,000
So the more negative images, the more malicious data we have in that set, the more the set will use, such as, you know, glue on pizza as another example.

129
00:20:34,000 --> 00:20:38,000
I think we will chat about that in a bit as well. But, you know, the silly things like that.

130
00:20:38,000 --> 00:20:48,000
And then an additional thing is what happens if you put unintentionally put that data in or bad quote unquote bad, like, sensitive data.

131
00:20:48,000 --> 00:21:03,000
So there are attacks on the models themselves called model and version attacks where using a set of keywords, grab personally identified for personally identifiable information

132
00:21:03,000 --> 00:21:10,000
from that data, which in a normal basis, you may not be able to get that data from the data set. Excuse me.

133
00:21:10,000 --> 00:21:19,000
And it's, you know, those are the three main attack vectors or forms of attacks. And I did.

134
00:21:19,000 --> 00:21:47,000
You know, in preparation for this, I did pull up a couple of actual attacks where in 2023, there was a some researchers discovered attack campaign against a vulnerability in Ray, one of the open source AI frameworks and pretty much the because there was a lack of separation

135
00:21:47,000 --> 00:21:55,000
and lack of data sterilization in the, in the framework and in that data set.

136
00:21:55,000 --> 00:22:09,000
The attackers were able to glean sensitive data work able to glean sensitive data from the data set as well as use the victims own processing power against them.

137
00:22:09,000 --> 00:22:28,000
So the attackers would come in, run their commands, run their queries and get all of this SPI from the customer without them knowing about it and using their stuff. So, you know, you might think old school hackers utilizing their own computers utilizing their own VMs.

138
00:22:28,000 --> 00:22:40,000
You know, all right me a worm. Spread it out, then maybe it'll get into the wild like Michelangelo or whoever, or whatever the happy birthday but from back in the day.

139
00:22:40,000 --> 00:22:45,000
But then, these days, you know, with this AI sets.

140
00:22:45,000 --> 00:23:05,000
You're having adversaries to be able to log into a computer and utilize the victims competing power and data set against them. So, it's, you know, kind of sketchy, and then another piece of it and towards the data poisoning side, and then the, I guess I'll just make an adjective out

141
00:23:05,000 --> 00:23:21,000
of this the Hitlerization of the data and of the model. There's what's called a skill attack, and Microsoft, the fund or sorry, Microsoft had funded some investigations some research is how to break AIs.

142
00:23:21,000 --> 00:23:42,000
And they they're, they're researching a few weeks ago. I think a month ago almost exactly a month ago, published details about a technique that bypasses all the guardrails used by the AI model, the AI model makers to prevent the generative chatbots from creating

143
00:23:42,000 --> 00:24:09,000
harmful content. And this, this technique is called skeleton key. And so, they these researchers were able to get, let's see what it says, the Metas Lama 3 instructions, Google Gemini Pro and Claude 3, Entropix Claude 3 into explaining how to make a molotov cocktail.

144
00:24:09,000 --> 00:24:29,000
So, that's something you would not, you know, it's well known. But, you know, we don't expect an AI chatbot to say, here's how to make a molotov cocktail. You ask a question, hey, how do I combine a rag, gasoline, and a lighter into making this into making a bomb?

145
00:24:29,000 --> 00:24:39,000
And normally the chatbot would not tell you here the ingredients here, the proportions, but these researchers were able to to do that with this model or these 3 models.

146
00:24:39,000 --> 00:24:58,000
And luckily, those vulnerabilities are being patched within the models. However, since it's a few data sets, since you have to sanitize all this data, even getting like, rolling the technology back to the point, rolling the data back to the point, you know,

147
00:24:58,000 --> 00:25:02,000
rolling the data back to the point where it's set, where it's sanitized again. Yeah.

148
00:25:02,000 --> 00:25:11,000
That could take weeks, months, but yeah, so, but, you know, that's kind of, yes, we have been hacked and yeah.

149
00:25:11,000 --> 00:25:20,000
And it's just going to continue to happen. I think we'll just see more and more of certainly without the security on top of it. It's inevitable.

150
00:25:20,000 --> 00:25:33,000
So why, what are the reasons that AI is vulnerable? Joel, maybe I'll give that one to you. Why is AI vulnerable to attacks?

151
00:25:33,000 --> 00:25:42,000
Yeah, it's, you know, there's a lot of different kind of ways you can look at this and avenues as to why or how it's vulnerable.

152
00:25:42,000 --> 00:25:52,000
You know, Sudhir covered well a lot of the how, but the why, a lot of it boils down to the complexity and the black box nature of AI.

153
00:25:52,000 --> 00:26:12,000
So, you know, these algorithms and databases are so huge, take so long to train and cost so much that it's really, you know, no one really has a good grasp of what is happening in there to know quite what to protect against.

154
00:26:12,000 --> 00:26:27,000
And, you know, also because these data sets take so long to train and, you know, build these algorithms, you kind of have that sunk cost fallacy because we, it's a black box, we can't just say don't use that.

155
00:26:27,000 --> 00:26:40,000
And researchers are having to come up with a way of, you know, we can't spend $150,000 or, you know, that was I think, CHET GPT-2, you know, the costs on these current models are insane.

156
00:26:40,000 --> 00:26:55,000
But, you know, we can't start over from scratch. It takes too long. There's so much money. So researchers are having to come up with a way of, I asked, just learned about this this morning because it was something that concerned me, unlearn previous knowledge.

157
00:26:55,000 --> 00:27:16,000
So basically, you know, my question was, you train a medical LLM on everything in PubMed. It's, you know, great. It supposedly knows the whole current medical knowledge of humanity. But then paper studies are retracted, you know, because they got their data wrong or even they just lied.

158
00:27:16,000 --> 00:27:33,000
Well, I won't go into the politics of my favorite story there, but yeah, but if you have this giant black box that was trained on that study that you know is no longer true, you don't want to go and have to completely retrain this bot.

159
00:27:33,000 --> 00:27:49,000
So researchers are learning ways to basically erase just that part of the knowledge, or that's what they're trying to figure out. We don't know how to do that yet. And it makes things really vulnerable as to, you know, how to use it.

160
00:27:49,000 --> 00:28:11,000
But the other thing is, it's part of us being in this hype cycle, is that people just have this assumption that AIs are right, and that they can be trusted, and they don't even apply it to say just this model or this one is, hey, AIs are right, I can trust the results.

161
00:28:11,000 --> 00:28:21,000
Not only that, if they're all right, if this new model hits hugging face tomorrow, well, I can just load that in my customer facing program because AI is trustworthy.

162
00:28:21,000 --> 00:28:42,000
That's right. And it's not, it's really short cutting some of these things that we have learned about software security and a good development lifecycle for years. And, you know, people are just kind of bypassing that because there is this inherent trust right now and AI that, you know, we're going to painfully

163
00:28:42,000 --> 00:29:01,000
learn that it's not as trustworthy and infallible as a lot of people seem to think. The assumptions are dangerous. Like, Andres, you mentioned at the beginning of the show, like, similarly to the cloud, my data must be safe when it's in the cloud, you know, those types of assumptions.

164
00:29:01,000 --> 00:29:15,000
Same thing with the assumptions with AI, you know, I was looking up a feature when ChatGPG just came out about Cisco ASAs when this particular feature came out and I asked it when and what version this feature was available.

165
00:29:15,000 --> 00:29:25,000
It came back with something that was so perfect and I was like, wow, it's exactly what I look for. And then I thought about it and I was like, that's impossible. That can't be the right version.

166
00:29:25,000 --> 00:29:33,000
Found the right documentation and it was wrong, but it certainly had me believe that, hey, that it came off as certainly correct.

167
00:29:33,000 --> 00:29:44,000
And I'll say that with that example that you just mentioned, imagine, you know, you've been doing firewalls for a long time. Imagine if it's somebody that just started with it.

168
00:29:44,000 --> 00:29:45,000
Yeah, right.

169
00:29:45,000 --> 00:29:47,000
Just believing it. It's, it's.

170
00:29:47,000 --> 00:29:59,000
Yeah, and to Joel, to your point, then you make that customer facing and you got a whole bunch of people relying on this data and everyone's assuming this is accurate and it's dangerous.

171
00:29:59,000 --> 00:30:02,000
Yeah.

172
00:30:02,000 --> 00:30:05,000
Good, good conversation.

173
00:30:05,000 --> 00:30:23,000
Now, just to kind of wrap some of that information up from from Cisco's point of view, if you don't mind, Joel, going over some of the things that we're doing in terms of using AI in our products.

174
00:30:23,000 --> 00:30:27,000
Do you have some examples that you can talk about?

175
00:30:27,000 --> 00:30:36,000
Yeah, absolutely. So, you know, Cisco uses AI all over the place, a lot internal and then a lot external.

176
00:30:36,000 --> 00:30:47,000
You know, if you have opened a TAC case in the last probably three and a half years, you might have gotten a note from Sherlock, who, yes, Sherlock Holmes works in Cisco TAC.

177
00:30:47,000 --> 00:30:48,000
That is.

178
00:30:48,000 --> 00:30:50,000
I thought it was a real person, man.

179
00:30:50,000 --> 00:30:57,000
I have customers. Yeah, like, hey, you've got a coworker named Sherlock Holmes. Isn't that funny?

180
00:30:57,000 --> 00:31:10,000
Yeah, we picked that name. Yeah, that's a bot who helps with our, you know, cases just to make sure for both Cisco and the customer that cases are moving forward in a timely manner.

181
00:31:10,000 --> 00:31:20,000
The bot can often help give the customer tips to help get things done quicker. So that's probably our biggest inside to outside facing thing.

182
00:31:20,000 --> 00:31:27,000
But then there are so many things that we do with AI that you may never even notice happen.

183
00:31:27,000 --> 00:31:39,000
So I would bet that you have not heard one time on this call the fact that I have a hundred pound dog who's been barking off and on behind that blurry white door behind me.

184
00:31:39,000 --> 00:31:46,000
That was the Babel Labs acquisition that is predictive AI that, you know, it senses the start of the bark.

185
00:31:46,000 --> 00:31:57,000
It knows the typical frequencies a dog bark is going to be at and it blocks that out of the audio feed coming from my device.

186
00:31:57,000 --> 00:32:17,000
You know, we have AI RRM, which automatically adjusts your wireless network, you know, changes channel with changes channels, everything just to make sure that the best coverage possible is happening with the wireless eight year company kind of, you know, towards the whole.

187
00:32:17,000 --> 00:32:28,000
Hey, we need to keep an eye on AI things. I will say that, you know, like with Sherlock, anything it didn't know the answer to that got kicked over to a human.

188
00:32:28,000 --> 00:32:33,000
It didn't try and guess an answer to a customer interaction.

189
00:32:33,000 --> 00:32:47,000
We had human oversight and similarly with AI RRM, you might not want your APs changing channels and doing things that would conflict with user experience during the day.

190
00:32:47,000 --> 00:32:51,000
So you can set a window, say, hey, eight to five, don't make any changes.

191
00:32:51,000 --> 00:32:55,000
Still pay attention to the wireless network that whole time.

192
00:32:55,000 --> 00:33:03,000
Come up with what you think would be good changes, but don't do that until out of business hours and then the next day, start all over again.

193
00:33:03,000 --> 00:33:06,000
You know, you've got those changes in place.

194
00:33:06,000 --> 00:33:10,000
Watch monitor everything again and come up with the next round.

195
00:33:10,000 --> 00:33:25,000
And that's, you know, I really like that towards the kind of responsible AI idea that, hey, we're going to let AI do things, make our lives better, but there is still the human control of we're not just handing over the keys.

196
00:33:25,000 --> 00:33:36,000
You know, we realize that things can be, you know, a negative impact. So let's still have a human touch within that whole experience.

197
00:33:36,000 --> 00:33:45,000
I've been seeing that point a lot recently as products evolve as well. Here's what we're recommending and this is fundamentally what will happen.

198
00:33:45,000 --> 00:33:52,000
But I'm going to allow the human to review this, apply it into action. Cisco HyperShield is an example of that work.

199
00:33:52,000 --> 00:34:01,000
This is what we're recommending. It even will say things like an Amazon review will do. This many other people have done this. This is what it looks like.

200
00:34:01,000 --> 00:34:12,000
This is my recommendation. You take it if you want to. Yeah. SD-WAN Analytics does the same thing. So we do our AI analysis of your Internet providers, your regular traffic.

201
00:34:12,000 --> 00:34:21,000
We might see that latency normally goes up at 2 p.m. on this link over here. But it's, hey, we notice this behavior.

202
00:34:21,000 --> 00:34:38,000
Our AI thinks that making this change will improve your experience. But you need to click apply over here first. We're not going to just do it without any input because that's always the fun of troubleshooting that black box of I don't.

203
00:34:38,000 --> 00:34:48,000
We're not just letting you have the key, but we're telling you why, why we suggest this change. And that really is kind of the difference between.

204
00:34:48,000 --> 00:34:57,000
There's, you know, there are technologies where autonomous AI makes sense. Self-driving cars. You can't wait for that human input.

205
00:34:57,000 --> 00:35:10,000
If the biker comes into the road, not that that went well. But yeah, there is so many times where, you know, we don't need to just trust AI and hand over the keys.

206
00:35:10,000 --> 00:35:19,000
There's still good reasons that a lot of this should still have the kind of that final say of, you know, let the human decide.

207
00:35:19,000 --> 00:35:33,000
Heck, that's why we have change review boards, even determining what humans do before someone changes, deploys an update to software. And, you know, that might cause an issue.

208
00:35:33,000 --> 00:35:42,000
I like that it's kind of when you use it in that way, it's bring your good ideas to the table being bring the good ideas to the table. I'll take a look at them and.

209
00:35:42,000 --> 00:35:47,000
You know, make sure that that is truly what I want.

210
00:35:47,000 --> 00:35:53,000
I didn't know we had that on the wireless and as the one side, that's pretty cool. That is.

211
00:35:53,000 --> 00:36:06,000
All right, so let's see that kind of covers. Oh, from the Cisco perspective, how we're using AI, which that's that's some remarkable stuff there. The Sherlock Holmes one. I've certainly had many customers Joel. Yeah.

212
00:36:06,000 --> 00:36:12,000
That's crazy. Sherlock. Who's the Sherlock guy? He's so smart. So, dear, what about.

213
00:36:12,000 --> 00:36:20,000
Similar question, but what is Cisco doing to secure the AI portion? Not about the using of, but how are we making sure that.

214
00:36:20,000 --> 00:36:32,000
Our AI is not getting hacked. Well, we can, I guess part of it, we can think the people over and I guess they're still in building for five, the security and trust organization.

215
00:36:32,000 --> 00:36:37,000
They are managing within Cisco the security of AIs for the most part.

216
00:36:37,000 --> 00:36:51,000
We have lots of internal WebEx team spaces and just kind of going over and discussing the best way to use AI, how to train AI to be ethical.

217
00:36:51,000 --> 00:36:56,000
To make sure AI exists within the secure development lifecycle.

218
00:36:56,000 --> 00:37:14,000
And to that point, we even have a new white belt program. I don't know if any of you guys remember 510 years ago. Mike, I'm not sure if you remember that we had the little squishies because they wanted us to get the black belt program for the security of lifecycle in tax.

219
00:37:14,000 --> 00:37:27,000
And they're doing something very similar with AI and the same thing, securing that on the tax side, gamifying the security of the AI, which I think is great.

220
00:37:27,000 --> 00:37:29,000
We're also doing.

221
00:37:29,000 --> 00:37:35,000
Well, within all of this, all of the individual department with Cisco collab.

222
00:37:35,000 --> 00:37:38,000
Data center enterprise security.

223
00:37:38,000 --> 00:38:03,000
We're partnering with industry experts to get what is best of breed in the latest understanding of how to secure AIs, how to make sure that the data we're putting into the into the AI is reliable is useful to get the AI to generate the content that we want from it.

224
00:38:03,000 --> 00:38:15,000
This goes back to the beginning of what we were talking about. We're no longer just relying on the AI to understand what we say and look into its database for already crafted answers.

225
00:38:15,000 --> 00:38:22,000
We're looking to the AI to generate new data from what it already knows. So, you know.

226
00:38:22,000 --> 00:38:30,000
To make sure that the data we put in, because at the end of the day, the AI is only as good as its data.

227
00:38:30,000 --> 00:38:39,000
So, establishing frameworks of ethics to make sure that we are only asking certain questions, or if we ask questions that are off the rails.

228
00:38:39,000 --> 00:38:45,000
Yeah, I just said sorry I can't do that Dave or something, something similar.

229
00:38:45,000 --> 00:38:53,000
We're not opening doors in a, in a space capsule or anything crazy like that, but that it goes in the same way.

230
00:38:53,000 --> 00:39:10,000
Those guardrails are in place for a reason to protect us to make sure that, to make sure that the AI and the large language model is robust, so that it's a bit more resilient against attacks.

231
00:39:10,000 --> 00:39:20,000
So that we don't inadvertently put secure or sensitive, personally identifiable information into the AI.

232
00:39:20,000 --> 00:39:23,000
So, all of these things together.

233
00:39:23,000 --> 00:39:31,000
Hats off to the security and trust, but they're spearheading all of this and including.

234
00:39:31,000 --> 00:39:45,000
I wouldn't say luckily, but kindly involving the rest of Cisco in that conversation, making sure that all of us work together to secure AI and secure those large language models.

235
00:39:45,000 --> 00:39:50,000
Yeah, and they are large. I know Joel mentioned the other day about.

236
00:39:50,000 --> 00:39:56,000
What was it? Joel 405 billion parameters on the 1 engine or.

237
00:39:56,000 --> 00:40:11,000
Yeah, the latest Lama 3.1 engine has a 405 billion parameter model. I, I'm trying to think the rough math. I think that's something like 147 gigs of GPU space.

238
00:40:11,000 --> 00:40:17,000
So, it doesn't fit in anything and video shipping. Now you'd have to have multiple GPUs to run it.

239
00:40:17,000 --> 00:40:29,000
Hope to do to your point on this topic. I'm glad to hear that Cisco is doing its part to secure that. And I do remember those little squishies, but nothing beats that little snorty squishy, which I know you remember from.

240
00:40:29,000 --> 00:40:35,000
Yeah, there you go. There you go.

241
00:40:35,000 --> 00:40:41,000
I throw them away.

242
00:40:41,000 --> 00:40:46,000
I got that.

243
00:40:46,000 --> 00:40:48,000
Was that Debbie?

244
00:40:48,000 --> 00:40:52,000
That's Debbie.

245
00:40:52,000 --> 00:41:06,000
That's awesome. So, a real, a real quick time check. I know we don't have too much time, but wanted to ask you the last questions there and what our customers can do to make sure that their AI is safe.

246
00:41:06,000 --> 00:41:11,000
So, you know, as we mentioned at the beginning.

247
00:41:11,000 --> 00:41:14,000
Training making sure the AI can't become a Hitler.

248
00:41:14,000 --> 00:41:22,000
Making sure that the data is providing is clean, consistent.

249
00:41:22,000 --> 00:41:34,000
Free of as much bias as possible. So, to making sure it's safe and to ensure the safety of the responses from that large large language model and the AI behind it.

250
00:41:34,000 --> 00:41:45,000
Auditing going through testing similar to what I'm assuming is doing with us with our eyes, the bridge. I think we have where.

251
00:41:45,000 --> 00:41:49,000
It's going to put knowingly bad.

252
00:41:49,000 --> 00:41:55,000
Props, or it's going to enter knowingly bad prompt and expect good clean answers from them.

253
00:41:55,000 --> 00:41:59,000
So that going through analyses for bias and fairness testing.

254
00:41:59,000 --> 00:42:09,000
So that you can ensure that the answers you're continuing continually. Well, the answers that you're continuing to receive from that model.

255
00:42:09,000 --> 00:42:15,000
Line up with your ethics line up with acceptable use the mores of the society we live in.

256
00:42:15,000 --> 00:42:18,000
And then making sure also that.

257
00:42:18,000 --> 00:42:27,000
If we run a debug and ask the AI model, why it's providing this information, it can explain. This is what data I use to get this info.

258
00:42:27,000 --> 00:42:30,000
And if that means going back and.

259
00:42:30,000 --> 00:42:36,000
Changing the model change in the data being put into it. We can more easily do that because of those logs that we have.

260
00:42:36,000 --> 00:42:40,000
They are explaining why it made these decisions and why it got to the output.

261
00:42:40,000 --> 00:42:43,000
And then monitoring it.

262
00:42:43,000 --> 00:42:49,000
Prevention and postvention. I guess you could say, you know.

263
00:42:49,000 --> 00:42:52,000
That's 100% of it at that point.

264
00:42:52,000 --> 00:42:55,000
And it just making sure it stays safe.

265
00:42:55,000 --> 00:43:00,000
And then also making sure that standards are kept.

266
00:43:00,000 --> 00:43:06,000
So, every industry has a set of standards, what you can do, what you're allowed to do.

267
00:43:06,000 --> 00:43:11,000
Thinking of the simplest ones, electric wiring. If you have a 10 gauge wire.

268
00:43:11,000 --> 00:43:16,000
Maybe you don't want to be putting 200 amps through it. But on the flip side.

269
00:43:16,000 --> 00:43:21,000
If somebody who's newer, Joel was mentioning newer.

270
00:43:21,000 --> 00:43:25,000
Newer and field, newer experience.

271
00:43:25,000 --> 00:43:35,000
Tech engineers or help desk engineers may answer a question, but if they don't have that experience, they won't know the 200 apps is too much to put through a 10 gauge wire.

272
00:43:35,000 --> 00:43:42,000
Or some, let's say 20 gauge, because I know 10 gauge depending on the wire. Maybe it could want 200 apps. But.

273
00:43:42,000 --> 00:43:48,000
You know, it's it's kind of chicken in the egg in a way, but then monitoring.

274
00:43:48,000 --> 00:44:00,000
And making sure everything stays safe. Yeah. Excellent. Excellent. I'll take a live question just with the, maybe a 32nd answer to this one. But.

275
00:44:00,000 --> 00:44:04,000
Carlos in the audience just brought up a interesting point.

276
00:44:04,000 --> 00:44:11,000
And I'll just kind of read this off, but pull that up here. Carlos pointed out that AI.

277
00:44:11,000 --> 00:44:16,000
Is, you know, accounting for.

278
00:44:16,000 --> 00:44:28,000
2 to 2 up to almost 4% of global greenhouse gas emissions due to the energy it's consuming due to its data centers. And we've been talking about how much process power takes to do this.

279
00:44:28,000 --> 00:44:33,000
In terms of may and maybe the answer is just better oversight, but any any.

280
00:44:33,000 --> 00:44:40,000
Wait, maybe 32nd opinions on the future of AI with respect to sustainability.

281
00:44:40,000 --> 00:44:50,000
I'll take that really quick. Carlos. It is a great question. It is something that concerns me a lot. You know, I mentioned the Amazon nuclear data center.

282
00:44:50,000 --> 00:45:05,000
But it's it really is. I think the best answer to this is some of the newer things we're coming out with, like small language models. You don't need a 405 billion parameter large language model that, you know, used umpteen gigawatts to train.

283
00:45:05,000 --> 00:45:23,000
If it's just doing a chat bot for support for your local tire shop or something, you know, I think a lot of companies are really looking at. Hey, rather than training this AI bot on every piece of literature ever created by humanity and written time.

284
00:45:23,000 --> 00:45:44,000
Well, let's train something well taken internal use case. Let's train a bot on just pack cases, you know, Cisco products, just the tech cases, the configuration guides can be a much smaller model, much less power intensive to create in the first place and to run while customers are using it.

285
00:45:44,000 --> 00:46:00,000
It's, you know, a really big call for bespoke models that are industry specific or subject specific that allow us to rather than just take that sledgehammer of throwing all these GPUs and all this power at everything.

286
00:46:00,000 --> 00:46:19,000
Let's be wiser about it and you know more judicious and our use of these things because yeah we can't continue the way things are going with these new models, these giant new models every month or two, and not just from one company but from meta Google and

287
00:46:19,000 --> 00:46:30,000
anthropic, you know, Microsoft, et cetera. It's just it's it's not sustainable. Great, great. Really quickly.

288
00:46:30,000 --> 00:46:47,000
Since you since you mentioned, like, exceeding the greenhouse gas emissions of the airplane industry, or the aviation industry. In my opinion, it's going to go the same way where the planes at the beginning were super dirty until we figured out ways to make them more efficient.

289
00:46:47,000 --> 00:46:58,000
And maybe we don't fly routes over the North Pole as often, or things of that nature. And I think it'll be the same way Joel's mentioning the small English models.

290
00:46:58,000 --> 00:47:02,000
I think we're going to be very brazen in our disregard for the environment.

291
00:47:02,000 --> 00:47:13,000
As long as the money is there, and as long as people can pay for greenhouse gas or carbon credits, like stuff like that.

292
00:47:13,000 --> 00:47:25,000
Until we can regulate that. I think it'll just take regulation as with the aviation industry. It took regulation. I think it'll be the exact same thing with it with the future of AI.

293
00:47:25,000 --> 00:47:40,000
And maybe we can offset some of that with the advancements the AI does on the other side of things for resolving problems. Maybe AI consumes a certain amount of percentage of greenhouse gas, but it's also

294
00:47:40,000 --> 00:47:47,000
coming up with ideas to remove 10% of it. So great question. Thank you, Carlos. Appreciate the response there.

295
00:47:47,000 --> 00:47:55,000
All right, so got about 2 minutes left. Let's jump to the most important part of this call here, which would be the dad jokes here.

296
00:47:55,000 --> 00:48:03,000
I just see you right here, man. What you want to take the 1st 1.

297
00:48:03,000 --> 00:48:06,000
All that. All right. All right.

298
00:48:06,000 --> 00:48:17,000
Let's do this. This I know you can't tell on the screen here, but there are like, 200 people right over in this area here. So.

299
00:48:17,000 --> 00:48:24,000
They're all kind of excited to see this. I got I got a, I got a horrible and a wonderful dad joke for you.

300
00:48:24,000 --> 00:48:29,000
How does an AI get rid of its bad habits?

301
00:48:29,000 --> 00:48:35,000
I don't know.

302
00:48:35,000 --> 00:48:40,000
It learns a new algorithm.

303
00:48:40,000 --> 00:48:45,000
Was that the good 1 or the bad 1?

304
00:48:45,000 --> 00:48:53,000
By AI, you know, we're, we're, we're, yeah, exactly.

305
00:48:53,000 --> 00:48:59,000
You got another one.

306
00:48:59,000 --> 00:49:02,000
Sure. I'll go for it. I'll go for another 1.

307
00:49:02,000 --> 00:49:10,000
What's the, why did the computer go to art school?

308
00:49:10,000 --> 00:49:14,000
Anyone paint brushes.

309
00:49:14,000 --> 00:49:19,000
To learn how to draw a better conclusion.

310
00:49:19,000 --> 00:49:25,000
Nice. I like that. I like that.

311
00:49:25,000 --> 00:49:31,000
Why couldn't the little boy go to see the pirate movie?

312
00:49:31,000 --> 00:49:40,000
How about the rating? This 1 was rated R.

313
00:49:40,000 --> 00:49:45,000
They're all terrible.

314
00:49:45,000 --> 00:49:51,000
What do you got? I got this one.

315
00:49:51,000 --> 00:49:58,000
Why did the robot, the robot go in a diet?

316
00:49:58,000 --> 00:50:02,000
And it's because it had too many bites.

317
00:50:02,000 --> 00:50:05,000
All right. You win so far. Let's go.

318
00:50:05,000 --> 00:50:15,000
Let's see what you got.

319
00:50:15,000 --> 00:50:17,000
I'll wrap it up with the one. I think my only joke that I always use with my daughters. So knock, knock, Sidir.

320
00:50:17,000 --> 00:50:19,000
Who's there?

321
00:50:19,000 --> 00:50:21,000
Orange.

322
00:50:21,000 --> 00:50:23,000
Orange who?

323
00:50:23,000 --> 00:50:27,000
Orange glad I didn't say banana.

324
00:50:27,000 --> 00:50:35,000
I'll have my take on as well as Andre's procedure. Any closing thoughts on your end?

325
00:50:35,000 --> 00:50:39,000
Uh, I think I'm going to have to say orange.

326
00:50:39,000 --> 00:50:44,000
I'm going to have to say orange.

327
00:50:44,000 --> 00:50:48,000
I'm going to have to say orange.

328
00:50:48,000 --> 00:50:51,000
I'm going to have to say orange.

329
00:50:51,000 --> 00:50:57,000
I'm going to have to say orange.

330
00:50:57,000 --> 00:51:02,000
No, I got, I got nothing useful.

331
00:51:02,000 --> 00:51:08,000
I say that we, I just want to wait and see what we get within Cisco for AI going into everything.

332
00:51:08,000 --> 00:51:13,000
If we're successful the next fiscal year, that's a big one.

333
00:51:13,000 --> 00:51:15,840
next fiscal year, that's a big one for me.

334
00:51:15,840 --> 00:51:18,700
See how the market reacts and see how our competitors

335
00:51:18,700 --> 00:51:21,960
also have improved their AI capabilities.

336
00:51:23,760 --> 00:51:24,600
Very good, Joel?

337
00:51:26,720 --> 00:51:29,920
Just a couple things, don't be scared of AI,

338
00:51:29,920 --> 00:51:33,240
it is something new to all of us, we're all learning.

339
00:51:33,240 --> 00:51:35,040
Kind of step into it.

340
00:51:35,040 --> 00:51:37,920
It's, we're going to have to learn new ways

341
00:51:37,920 --> 00:51:40,360
to secure things and new attack factors,

342
00:51:40,360 --> 00:51:43,920
but it's, heck, that's why I like technology,

343
00:51:43,920 --> 00:51:45,280
it's something new.

344
00:51:45,280 --> 00:51:47,160
But while you're doing that, don't forget,

345
00:51:47,160 --> 00:51:49,680
it's still software, you still need to do

346
00:51:49,680 --> 00:51:52,000
all the irregular things you've always done

347
00:51:52,000 --> 00:51:55,960
to secure applications running in your network.

348
00:51:57,800 --> 00:51:59,960
Great, great, well, good conversation today.

349
00:51:59,960 --> 00:52:02,640
I like how we kind of started off with talking about

350
00:52:02,640 --> 00:52:04,360
just generally how AI works.

351
00:52:04,360 --> 00:52:06,520
I know Joel, you spoke to that, the components,

352
00:52:06,520 --> 00:52:07,860
where the data lives.

353
00:52:07,860 --> 00:52:12,020
I do think a great question Andres had was,

354
00:52:12,020 --> 00:52:13,660
how is it that we seem to have all of a sudden

355
00:52:13,660 --> 00:52:17,540
just heard about AI, but as we saw Sudhir kind of answer,

356
00:52:17,540 --> 00:52:19,820
no, it actually has been around a while,

357
00:52:19,820 --> 00:52:22,840
but we talked about having the resources

358
00:52:22,840 --> 00:52:26,060
to really see it explode and to see the benefits of it.

359
00:52:26,060 --> 00:52:29,380
And we got into the dangers of AI,

360
00:52:29,380 --> 00:52:33,260
the Skynet terminator stuff, and maybe that'll be

361
00:52:33,260 --> 00:52:36,620
episode two of our AI call when that does happen.

362
00:52:36,620 --> 00:52:39,660
But on a serious note, people's data getting exposed,

363
00:52:39,660 --> 00:52:42,260
or Joel, you brought up an interesting point

364
00:52:42,260 --> 00:52:44,580
about just giving out wrong information

365
00:52:44,580 --> 00:52:46,580
and having people run with that information,

366
00:52:46,580 --> 00:52:50,300
not having your models get updated as we evolve

367
00:52:50,300 --> 00:52:52,420
and you have that medical example.

368
00:52:53,940 --> 00:52:56,100
Sudhir, you brought up a couple hacks, interesting.

369
00:52:56,100 --> 00:52:57,220
I think we'll see more of those,

370
00:52:57,220 --> 00:53:02,220
especially as security just gets left out of AI, unfortunately.

371
00:53:03,380 --> 00:53:05,540
So hopefully that won't happen,

372
00:53:05,540 --> 00:53:07,580
but we all know that it will.

373
00:53:07,580 --> 00:53:09,580
Andres, over to you.

374
00:53:09,580 --> 00:53:11,340
Yeah, no, and that's a good point, Mike.

375
00:53:11,340 --> 00:53:14,660
I know, for example, in those examples you mentioned,

376
00:53:14,660 --> 00:53:16,340
and the ones that Sudhir mentioned,

377
00:53:16,340 --> 00:53:19,020
we're gonna put them on the episode notes as well.

378
00:53:19,020 --> 00:53:24,020
But just understanding, and I think we all knew this one,

379
00:53:24,860 --> 00:53:27,700
AI is vulnerable to attacks.

380
00:53:27,700 --> 00:53:31,020
We've seen a lot of stuff going on out there.

381
00:53:31,020 --> 00:53:35,140
So just making sure, I really like this

382
00:53:35,140 --> 00:53:37,300
because it opened up to other things

383
00:53:37,300 --> 00:53:38,780
that we probably didn't know.

384
00:53:38,780 --> 00:53:41,940
In my case, I didn't know, so it helps a lot.

385
00:53:43,740 --> 00:53:48,220
What are we doing from the Cisco's point of view?

386
00:53:48,220 --> 00:53:52,300
What are we doing as far as AI into our products?

387
00:53:52,300 --> 00:53:53,520
How we're securing it?

388
00:53:54,460 --> 00:53:56,300
Sounds responsible to me.

389
00:53:56,300 --> 00:53:58,820
Of course, anything can happen.

390
00:53:58,820 --> 00:54:03,100
So like Sudhir mentioned, just wait and see

391
00:54:03,100 --> 00:54:05,140
what we can accomplish with AI

392
00:54:05,140 --> 00:54:07,580
and where else we're gonna see it.

393
00:54:09,100 --> 00:54:11,260
The hype just brings a lot of opportunity

394
00:54:11,260 --> 00:54:14,940
and a lot of moving fast, breaking things.

395
00:54:14,940 --> 00:54:19,100
So hopefully we don't break too many things, if anything.

396
00:54:19,100 --> 00:54:21,060
Okay, fingers crossed.

397
00:54:21,060 --> 00:54:24,540
Yeah, and then the last thing is

398
00:54:24,540 --> 00:54:26,100
recommendation for our customers.

399
00:54:26,100 --> 00:54:27,940
Like if you're getting into AI,

400
00:54:27,940 --> 00:54:29,620
if you're building something,

401
00:54:29,620 --> 00:54:33,260
make sure you keep a few things in check.

402
00:54:33,260 --> 00:54:34,780
Make sure you monitor it.

403
00:54:34,780 --> 00:54:36,780
Make sure you test it out all the time

404
00:54:36,780 --> 00:54:39,900
and make sure that you keep it secure.

405
00:54:39,900 --> 00:54:42,580
And that's all I have from the episode.

406
00:54:42,580 --> 00:54:43,500
That was great.

407
00:54:43,500 --> 00:54:46,220
Great, I really enjoyed today's conversation.

408
00:54:46,220 --> 00:54:48,020
Sudhir and Joel, thanks for taking the time

409
00:54:48,020 --> 00:54:49,260
to join us on the show today.

410
00:54:49,260 --> 00:54:51,340
You guys have a wealth of experience

411
00:54:51,340 --> 00:54:54,060
pertaining specifically to AIs.

412
00:54:54,060 --> 00:54:55,380
So really have enjoyed it.

413
00:54:55,380 --> 00:54:57,700
And thanks for all the good you do in the world.

414
00:54:57,700 --> 00:55:00,100
Keep in doing your part to keep us secure.

415
00:55:01,220 --> 00:55:02,820
Andres, our next two episodes

416
00:55:02,820 --> 00:55:05,260
kinda are a pairing going together.

417
00:55:05,260 --> 00:55:08,780
September 5th is a deep dive on XDR.

418
00:55:09,780 --> 00:55:11,260
And kind of a special surprise,

419
00:55:11,260 --> 00:55:13,420
the following day, September 6th,

420
00:55:13,420 --> 00:55:16,100
we're gonna see a live demo of Cisco XDR.

421
00:55:16,100 --> 00:55:17,900
And so get to see the dashboard.

422
00:55:17,900 --> 00:55:19,420
That'll be pretty cool.

423
00:55:19,420 --> 00:55:21,700
We'll send out some updated invites for that.

424
00:55:21,700 --> 00:55:24,580
I really enjoyed today's conversation, guys.

425
00:55:24,580 --> 00:55:27,940
Please stay secure and we'll see you on the next show.

426
00:55:27,940 --> 00:55:29,140
Bye everybody.

427
00:55:29,140 --> 00:55:54,980
Thank you.