1 00:00:00,000 --> 00:00:09,600 Welcome to the Azure Security Podcast, where we discuss topics relating to security, privacy, 2 00:00:09,600 --> 00:00:13,280 reliability and compliance on the Microsoft Cloud Platform. 3 00:00:13,280 --> 00:00:17,820 Hey everybody, welcome to episode 108. 4 00:00:17,820 --> 00:00:19,360 This week we actually have the full gang. 5 00:00:19,360 --> 00:00:22,580 It's myself, Michael, Mark, Sarah and Gladys. 6 00:00:22,580 --> 00:00:28,200 And our guest this week is Diana, who's here to talk to us about securing Gen.ai applications. 7 00:00:28,200 --> 00:00:31,400 But before we get to our guests, let's take a little lap around the news. 8 00:00:31,400 --> 00:00:33,280 Gladys, why don't you kick things off. 9 00:00:33,280 --> 00:00:35,800 Hi everyone, happy new year. 10 00:00:35,800 --> 00:00:40,600 I only have two set of news that I want to discuss about. 11 00:00:40,600 --> 00:00:44,200 The first one is about TLS 1.3. 12 00:00:44,200 --> 00:00:53,240 Starting March 1st, 2025, Azure services would require using TLS 1.2 or higher in the support 13 00:00:53,240 --> 00:00:56,360 for TLS 1.0 and 1.1. 14 00:00:56,360 --> 00:01:02,520 This is to enhance the security and provide best-in-class encryption. 15 00:01:02,520 --> 00:01:10,540 So all the services will enable 1.3 on the same day and disable the earlier support. 16 00:01:10,540 --> 00:01:17,160 The next news that I wanted to talk about is several features that Microsoft Defender 17 00:01:17,160 --> 00:01:20,040 for Cloud released in December. 18 00:01:20,040 --> 00:01:25,960 The first one is that they're changing the scanning turbo for AWS and GCP. 19 00:01:25,960 --> 00:01:35,320 The scanning turbo settings, which is capturing all the security findings from AWS and GCP, 20 00:01:35,320 --> 00:01:41,480 determines how often Defender for Cloud discover that information. 21 00:01:41,480 --> 00:01:49,000 So this change ensured a more balanced scanning process and it optimized performance and minimized 22 00:01:49,000 --> 00:01:52,520 the risk of reaching API limits. 23 00:01:52,520 --> 00:02:01,680 The next feature that Microsoft Defender changed is the way Microsoft Defender for endpoint 24 00:02:01,680 --> 00:02:08,080 will be required in order to receive file integrity monitoring experience. 25 00:02:08,080 --> 00:02:15,520 And basically just some of the data provided provide better monitoring information. 26 00:02:15,520 --> 00:02:22,040 And last, Defender for Cloud security posture management, sensitive scanning capability 27 00:02:22,040 --> 00:02:28,760 will be included in Azure file shares in addition to blog containers. 28 00:02:28,760 --> 00:02:35,040 So in my part of the world, I've gotten a little bit ranty on LinkedIn and the other 29 00:02:35,040 --> 00:02:36,520 social networks. 30 00:02:36,520 --> 00:02:41,160 Professional, positive, encouraging, but still like kind of really ranting about a couple 31 00:02:41,160 --> 00:02:42,160 different things. 32 00:02:42,160 --> 00:02:47,920 I've become more and more convicted lately that ultimately there's this sort of simple 33 00:02:47,920 --> 00:02:52,200 principle in life that you can't blame someone for decisions or actions they don't make. 34 00:02:52,200 --> 00:02:56,220 And the reality is that's often what happens in security. 35 00:02:56,220 --> 00:03:00,760 And so one of the things I've gotten a lot of clarity on and was discussing publicly 36 00:03:00,760 --> 00:03:07,200 in the socials was just around, ultimately, the way that organizations treat security 37 00:03:07,200 --> 00:03:09,880 is often kind of broken. 38 00:03:09,880 --> 00:03:13,440 And so it's like, hey, you don't blame the lawyer if they said, hey, that's a bad idea. 39 00:03:13,440 --> 00:03:17,760 Or you don't blame the finance person if the CEO made the decision to spend a whole 40 00:03:17,760 --> 00:03:20,040 bunch of money on something that didn't make sense. 41 00:03:20,040 --> 00:03:23,120 You know, it's just, you know, but yeah, we do that in security. 42 00:03:23,120 --> 00:03:25,880 We say, oh, you know, they said there's no maintenance windows. 43 00:03:25,880 --> 00:03:26,880 They didn't patch it. 44 00:03:26,880 --> 00:03:27,880 They didn't configure it. 45 00:03:27,880 --> 00:03:30,120 They didn't do any app sec or threat modeling or anything. 46 00:03:30,120 --> 00:03:33,020 And we're going to blame the security team who had nothing to do with any of the production 47 00:03:33,020 --> 00:03:36,620 of all those things and blame them for the incident. 48 00:03:36,620 --> 00:03:41,440 And so it's this really interesting kind of oddity that we deal with in security where 49 00:03:41,440 --> 00:03:47,000 it just hasn't been seen as sort of like a normal support discipline like legal or HR 50 00:03:47,000 --> 00:03:51,720 or finance or any of that kind of stuff in an organization. 51 00:03:51,720 --> 00:03:55,260 And it really, you know, it's one of the biggest changes, I think, that we need to make as 52 00:03:55,260 --> 00:03:56,840 an industry. 53 00:03:56,840 --> 00:03:59,160 So I'll pop a couple links to some of these discussions. 54 00:03:59,160 --> 00:04:01,240 The other one is just around collaboration. 55 00:04:01,240 --> 00:04:04,160 It's really important that, you know, you look at security as something you do, not 56 00:04:04,160 --> 00:04:05,160 something you have. 57 00:04:05,160 --> 00:04:10,960 And it's just you're constantly taking inputs from either incidents or other events inside 58 00:04:10,960 --> 00:04:17,960 and outside the organization and using those as a stimulus to drive security improvements 59 00:04:17,960 --> 00:04:20,160 and run your processes. 60 00:04:20,160 --> 00:04:22,680 So got a couple links to those there. 61 00:04:22,680 --> 00:04:25,120 And then not quite released yet. 62 00:04:25,120 --> 00:04:28,000 So think of this as more of a tease than an actual announce. 63 00:04:28,000 --> 00:04:36,840 But putting the very final touches on the updated CISO workshop and the updated cybersecurity 64 00:04:36,840 --> 00:04:37,960 reference architecture. 65 00:04:37,960 --> 00:04:44,000 So those are getting close to a release and hopefully we'll have some good news in future 66 00:04:44,000 --> 00:04:45,760 episodes very soon. 67 00:04:45,760 --> 00:04:46,760 That's all I got. 68 00:04:46,760 --> 00:04:51,640 Okay, before we get into my news, I just want to add a bit more detail about the TLS 1.3 69 00:04:51,640 --> 00:04:53,560 item that Gladys mentioned. 70 00:04:53,560 --> 00:04:57,880 Correctly she said that, you know, TLS 1.3 provides better encryption. 71 00:04:57,880 --> 00:05:02,640 That is true, better encryption, and it's also a stronger integrity check than prior 72 00:05:02,640 --> 00:05:04,080 versions of TLS. 73 00:05:04,080 --> 00:05:07,720 But honestly, the most important thing about TLS 1.3 is actually the strongest server authentication 74 00:05:07,720 --> 00:05:08,720 that you get. 75 00:05:08,720 --> 00:05:11,120 And you get that via better cipher suites. 76 00:05:11,120 --> 00:05:16,080 In TLS 1.3, you have to realize that one of the big things that the protocol did is basically 77 00:05:16,080 --> 00:05:21,840 got rid of a whole bunch of legacy and pretty lousy authentication ciphers. 78 00:05:21,840 --> 00:05:22,840 So they've all gone. 79 00:05:22,840 --> 00:05:26,660 So TLS 1.3 is a lot more streamlined than TLS 1.2 and prior versions. 80 00:05:26,660 --> 00:05:31,040 So wherever possible, you should always be going to TLS 1.3 rather than TLS 1.2, even 81 00:05:31,040 --> 00:05:34,880 though we do support TLS 1.2 simply because of backward compatibility. 82 00:05:34,880 --> 00:05:40,080 With that being said, we are seeing compat problems with TLS 1.0 and 1.1 being deprecated. 83 00:05:40,080 --> 00:05:44,160 But honestly, they are so old and so broken that you really should be moving to 1.2 and 84 00:05:44,160 --> 00:05:45,160 preferably 1.3. 85 00:05:45,160 --> 00:05:47,160 All right, so here's my news. 86 00:05:47,160 --> 00:05:50,720 So the first one is from my old stomping ground. 87 00:05:50,720 --> 00:05:54,720 SQL managed instance now supports service endpoints for Azure storage, which basically 88 00:05:54,720 --> 00:06:00,740 means that you can restrict where SQL managed instances pull their data from. 89 00:06:00,740 --> 00:06:05,360 This lets you really have really fine-grained control over the storage accounts that are 90 00:06:05,360 --> 00:06:06,360 used. 91 00:06:06,360 --> 00:06:11,000 Net effect is stronger security and honestly better egress control, so reducing the chance 92 00:06:11,000 --> 00:06:13,920 that attackers can actually egress data. 93 00:06:13,920 --> 00:06:21,460 Next one is Azure confidential ledger now has achieved SOC 2 type 2 compliance, a big 94 00:06:21,460 --> 00:06:24,740 deal for all those who require SOC 2 compliance. 95 00:06:24,740 --> 00:06:29,320 And the last one that I have is again from my old stomping ground, Azure database for 96 00:06:29,320 --> 00:06:35,000 MySQL now supports accelerated logs by default, including with support for customer managed 97 00:06:35,000 --> 00:06:36,600 keys. 98 00:06:36,600 --> 00:06:40,240 Accelerated logs are not what we know in the security industry is like log files in terms 99 00:06:40,240 --> 00:06:43,240 of audit logs and debug logs and so on. 100 00:06:43,240 --> 00:06:45,160 These are actually the transaction logs. 101 00:06:45,160 --> 00:06:50,200 Not only does it support accelerated logs by default, but it also supports customer 102 00:06:50,200 --> 00:06:52,440 managed keys as well, which is really great to see. 103 00:06:52,440 --> 00:06:54,820 In fact, that's the only reason I brought it up is because it does support customer 104 00:06:54,820 --> 00:06:55,820 managed keys as well. 105 00:06:55,820 --> 00:07:00,640 So you get a huge performance boost and your log files can still be protected by your own 106 00:07:00,640 --> 00:07:01,640 keys. 107 00:07:01,640 --> 00:07:04,920 All right, so with that, let's switch our attention now to our guest. 108 00:07:04,920 --> 00:07:09,120 As I mentioned before, our guest this week is Diana and she's here to talk to us about 109 00:07:09,120 --> 00:07:10,700 securing Gen.ai app. 110 00:07:10,700 --> 00:07:12,960 So Diana, thank you so much for joining us this week. 111 00:07:12,960 --> 00:07:16,840 Would you like to take a moment and sort of introduce yourself to our listeners? 112 00:07:16,840 --> 00:07:17,840 Sounds good. 113 00:07:17,840 --> 00:07:18,840 Hi everyone. 114 00:07:18,840 --> 00:07:23,120 Thank you so, so much for having me here today. 115 00:07:23,120 --> 00:07:25,000 My name is Diana Misesad. 116 00:07:25,000 --> 00:07:31,800 I am a product manager at Microsoft on the Entra Copilot team focused on bringing Gen.ai 117 00:07:31,800 --> 00:07:33,520 capabilities to Entra. 118 00:07:33,520 --> 00:07:38,240 So you might have heard of our recent public preview release of security copilot embedded 119 00:07:38,240 --> 00:07:41,000 in Entra last November. 120 00:07:41,000 --> 00:07:44,840 So super excited to be here today to talk about a very important topic. 121 00:07:44,840 --> 00:07:48,560 All right, so this one's actually a very interesting episode because it's actually going to be 122 00:07:48,560 --> 00:07:53,020 one of four episodes covering securing Gen.ai applications. 123 00:07:53,020 --> 00:07:57,440 So before we sort of get into the guts of this, Diana, why don't you spend just a quick 124 00:07:57,440 --> 00:08:02,500 brief moment and go over why we've got four episodes and just really quickly touch on 125 00:08:02,500 --> 00:08:05,220 what the other episodes are so people know what to expect. 126 00:08:05,220 --> 00:08:13,280 So with a group of colleagues on the identity space, we wrote a MS Learn article on how 127 00:08:13,280 --> 00:08:18,160 you can use Entra capabilities to secure your Gen.ai apps. 128 00:08:18,160 --> 00:08:23,280 So the reason why we're going to have a series of four episodes coming up on this topic is 129 00:08:23,280 --> 00:08:28,200 because we do want to take the time to talk about each of the very important steps and 130 00:08:28,200 --> 00:08:33,320 each of the very important sections of this article where we're going to share more about 131 00:08:33,320 --> 00:08:38,040 how you can use Microsoft Entra capabilities to protect your environment, to protect your 132 00:08:38,040 --> 00:08:39,040 Gen.ai apps. 133 00:08:39,040 --> 00:08:45,320 And this is the first episode of this series where we're going to be talking about why 134 00:08:45,320 --> 00:08:52,800 it's important to secure your Gen.ai apps, why businesses are implementing AI more than 135 00:08:52,800 --> 00:08:57,840 ever and how you can learn more about this because I think it's a topic that everyone 136 00:08:57,840 --> 00:09:00,280 should know a little bit about, right? 137 00:09:00,280 --> 00:09:03,480 Everyone should be able to understand the impact. 138 00:09:03,480 --> 00:09:05,800 Everyone should be able to understand the importance of these topics. 139 00:09:05,800 --> 00:09:11,880 So super excited to kick off this series of four episodes and we hope that everyone enjoys 140 00:09:11,880 --> 00:09:13,240 and learns a lot. 141 00:09:13,240 --> 00:09:18,080 So most people have heard about AI, actually, hopefully everyone, right? 142 00:09:18,080 --> 00:09:23,720 Even my kids keep talking about AI and how to use it in school and things like that. 143 00:09:23,720 --> 00:09:33,680 But can you explain a little bit about why it's important to have AI in the side of business 144 00:09:33,680 --> 00:09:35,680 today and secure it? 145 00:09:35,680 --> 00:09:38,160 Yeah, that's a great question. 146 00:09:38,160 --> 00:09:41,160 I always like to say that that's everything in life. 147 00:09:41,160 --> 00:09:44,440 It's always a good and a not so good side of everything. 148 00:09:44,440 --> 00:09:47,400 And with AI, it's the same thing, right? 149 00:09:47,400 --> 00:09:54,360 We've seen how AI is so helpful for multiple reasons in order to do multiple types of tasks. 150 00:09:54,360 --> 00:09:59,280 And many of the organizations that are implementing AI in their operations, they do feel some 151 00:09:59,280 --> 00:10:04,760 sort of anxiety about using AI because they're worried about sensitive data, the leakage 152 00:10:04,760 --> 00:10:07,040 of sensitive data, right? 153 00:10:07,040 --> 00:10:14,240 So being able to understand some of the risks that are associated with AI, using AI in your 154 00:10:14,240 --> 00:10:16,480 operations is so, so important. 155 00:10:16,480 --> 00:10:19,160 It's more important than ever, right? 156 00:10:19,160 --> 00:10:25,080 And in order to provide a better picture of this, I like to talk about AI discovery. 157 00:10:25,080 --> 00:10:30,040 And the way that I think about AI discovery is with every new piece of technology, you 158 00:10:30,040 --> 00:10:34,120 know, it's AI is this exciting new thing that everyone wants to learn about, everyone wants 159 00:10:34,120 --> 00:10:38,280 to play with AI and AI discovery is great, right? 160 00:10:38,280 --> 00:10:43,160 We can use AI for so many different reasons. 161 00:10:43,160 --> 00:10:47,440 For example, you know, if we have a PowerPoint presentation that we're working on with a 162 00:10:47,440 --> 00:10:52,640 colleague and we just lost track of it, I can use an AI assistant and say, hey, you 163 00:10:52,640 --> 00:10:56,800 know, can you show me what are some of the files that I've been working on with this 164 00:10:56,800 --> 00:10:57,800 colleague? 165 00:10:57,800 --> 00:11:01,200 An AI assistant can lead me to those files. 166 00:11:01,200 --> 00:11:02,200 And that's great, right? 167 00:11:02,200 --> 00:11:06,360 Because it really facilitates a lot of the things that we do in our day to day life and 168 00:11:06,360 --> 00:11:07,360 at work. 169 00:11:07,360 --> 00:11:10,960 And AI discovery is great, as I said, right? 170 00:11:10,960 --> 00:11:16,880 And if I am a bad actor, you know, and this is really important to understand, using AI, 171 00:11:16,880 --> 00:11:23,520 AI is also a great piece of technology, because if I'm a bad actor, I can get control of an 172 00:11:23,520 --> 00:11:24,520 identity. 173 00:11:24,520 --> 00:11:25,520 And this is very serious. 174 00:11:25,520 --> 00:11:29,880 And it's way more than just asking, hey, who I am, you know, if I can ask an AI assistant 175 00:11:29,880 --> 00:11:35,320 about all sorts of confidential information that I shouldn't have access to, that I shouldn't 176 00:11:35,320 --> 00:11:36,480 know about, right? 177 00:11:36,480 --> 00:11:39,420 So it could be like a new product release and things of that nature. 178 00:11:39,420 --> 00:11:44,520 So we can really see how AI can really be a force for good for certain situations, but 179 00:11:44,520 --> 00:11:51,040 can really be used, you know, for malicious type of intent, right? 180 00:11:51,040 --> 00:11:55,500 So understanding that there is a good and a bad of these and how to mitigate the result, 181 00:11:55,500 --> 00:11:59,280 you know, the backside of it is really important. 182 00:11:59,280 --> 00:12:06,160 So tell me about what are some of the key security risks that are associated with integrating 183 00:12:06,160 --> 00:12:09,840 GenAI into business operations? 184 00:12:09,840 --> 00:12:14,520 What are the things that you're seeing on your side of the house that people are worried 185 00:12:14,520 --> 00:12:15,520 about? 186 00:12:15,520 --> 00:12:16,520 No, I love that question. 187 00:12:16,520 --> 00:12:20,640 And I talk a lot about this with my friends as well. 188 00:12:20,640 --> 00:12:24,840 Because I was talking about intentional AI discovery, right, which is great. 189 00:12:24,840 --> 00:12:28,840 But then there's also unintentional AI discovery. 190 00:12:28,840 --> 00:12:32,680 And I think that everyone should know about these organizations, so you have trainings 191 00:12:32,680 --> 00:12:37,320 for their teams to make sure that they understand these type of situations. 192 00:12:37,320 --> 00:12:41,240 So you know, discoverability of AI, right? 193 00:12:41,240 --> 00:12:46,600 We're talking about, you know, situations where AI systems can provide you with information 194 00:12:46,600 --> 00:12:49,560 that you need for your job, for example, right? 195 00:12:49,560 --> 00:12:54,800 But then we have unintentional AI discovery, which is situations where the AI system provides 196 00:12:54,800 --> 00:12:57,240 you with information that you shouldn't have access to it. 197 00:12:57,240 --> 00:13:01,120 I think I provided a quick example earlier. 198 00:13:01,120 --> 00:13:05,000 But in terms of, you know, some of the key security risks that I've seen, and I have 199 00:13:05,000 --> 00:13:08,960 stories I was talking to Bailey, who is one of my colleagues that is going to be part 200 00:13:08,960 --> 00:13:11,720 of the upcoming episodes on this topic. 201 00:13:11,720 --> 00:13:20,440 She was telling me about a story where she heard about this person at her workplace that 202 00:13:20,440 --> 00:13:27,200 was asking an AI assistant about some sort of information that she needed for her job, 203 00:13:27,200 --> 00:13:28,200 right? 204 00:13:28,200 --> 00:13:32,840 And then she was prompted with information that she realized she shouldn't have access 205 00:13:32,840 --> 00:13:35,080 to and she was like, okay, what is going on? 206 00:13:35,080 --> 00:13:38,960 So that was a clear example of what happens with unintentional AI discovery. 207 00:13:38,960 --> 00:13:42,220 It's not like someone wakes up one morning and is like, oh, you know, like, let me get 208 00:13:42,220 --> 00:13:44,200 access to sensitive confidential information. 209 00:13:44,200 --> 00:13:49,480 No, it's like literally somebody who is like being told, hey, use AI, it's going to help 210 00:13:49,480 --> 00:13:51,720 you for your task and you use it. 211 00:13:51,720 --> 00:13:58,600 And then all of a sudden, the H&AI app is not being well maintained or managed and you 212 00:13:58,600 --> 00:14:03,880 have access to multiple files that you shouldn't have access to and you come out across this 213 00:14:03,880 --> 00:14:05,280 type of data, right? 214 00:14:05,280 --> 00:14:11,240 So for we know that, you know, there are situations like inside a risk where you know that there 215 00:14:11,240 --> 00:14:14,760 are people that are trying to do bad things within your organization. 216 00:14:14,760 --> 00:14:19,720 So this could be like, for example, asking an AI assistant for your colleague's performance 217 00:14:19,720 --> 00:14:23,360 evaluation, which you know, you are their colleague, you shouldn't have access to that. 218 00:14:23,360 --> 00:14:24,960 You are their peer, right? 219 00:14:24,960 --> 00:14:28,920 You shouldn't be able to query that type of information with an AI assistant. 220 00:14:28,920 --> 00:14:34,080 If you are their manager, you know, you can ask about, you know, evaluations or salary 221 00:14:34,080 --> 00:14:39,000 information because you do have access to the data, to the information. 222 00:14:39,000 --> 00:14:44,200 But if you don't have access to that information, you shouldn't come across that with an answer 223 00:14:44,200 --> 00:14:46,360 from an AI assistant, right? 224 00:14:46,360 --> 00:14:50,160 So you know, really talking about those types of real life scenarios. 225 00:14:50,160 --> 00:14:56,200 Another scenario that I have is with healthcare data where you can come across PII of patients 226 00:14:56,200 --> 00:14:58,600 or employees. 227 00:14:58,600 --> 00:15:02,840 Another example that I heard recently was in finance, right? 228 00:15:02,840 --> 00:15:08,640 So if someone in finance, let's say a financial analyst is asking an internal AI assistant 229 00:15:08,640 --> 00:15:11,160 about like financial modeling. 230 00:15:11,160 --> 00:15:15,600 If you are working at a big bang and you are, you want to know about like, hey, I'm looking 231 00:15:15,600 --> 00:15:22,280 into, you know, what should I be looking at this year in terms of like financial forecasting 232 00:15:22,280 --> 00:15:24,160 for such and such industry? 233 00:15:24,160 --> 00:15:28,120 And then all of a sudden, you find out about merger and acquisition that on the other side 234 00:15:28,120 --> 00:15:32,320 of the firm, someone is working on and you shouldn't know about that, right? 235 00:15:32,320 --> 00:15:38,120 So those are examples of unintentional AI discovery and these are very common today. 236 00:15:38,120 --> 00:15:42,980 So how do we make sure that we prevent those situations from happening? 237 00:15:42,980 --> 00:15:47,440 That's definitely one of the things that are going to be discussed in the next few episodes 238 00:15:47,440 --> 00:15:50,960 with my colleagues, but really, you know, trying to make sure that people understand, 239 00:15:50,960 --> 00:15:53,400 hey, these are things that are happening today. 240 00:15:53,400 --> 00:15:58,080 And the reason why they're happening is not because of AI, right? 241 00:15:58,080 --> 00:16:00,480 It's because we're not doing the basics well. 242 00:16:00,480 --> 00:16:06,320 And when we say the basics, we're talking about why are these situations happening? 243 00:16:06,320 --> 00:16:10,720 You know, these examples that I mentioned, they're being caused by things like excessive 244 00:16:10,720 --> 00:16:11,720 access. 245 00:16:11,720 --> 00:16:15,560 So like people are having privilege that they should no longer have. 246 00:16:15,560 --> 00:16:21,120 And this is very common when like people are changing jobs within an organization and they 247 00:16:21,120 --> 00:16:24,360 still have the old permissions that they had before, right? 248 00:16:24,360 --> 00:16:27,600 And they don't need those anymore, but they still have those permissions. 249 00:16:27,600 --> 00:16:32,220 Could also be caused by, for example, improper DLP labeling, right? 250 00:16:32,220 --> 00:16:35,900 So we are not labeling our files the way we should. 251 00:16:35,900 --> 00:16:37,640 And it could also be like data isolation. 252 00:16:37,640 --> 00:16:40,240 So these are things that are the basics. 253 00:16:40,240 --> 00:16:42,840 We should be taking care of these basic things. 254 00:16:42,840 --> 00:16:47,560 So one thing that we really want to highlight is like AI is not bad for your organization. 255 00:16:47,560 --> 00:16:50,240 AI is not bad for your operations. 256 00:16:50,240 --> 00:16:55,380 The reason why these situations keep happening is because we're not doing the basics well. 257 00:16:55,380 --> 00:16:59,960 So yeah, that's pretty much some of the examples, real life examples that I have for you all. 258 00:16:59,960 --> 00:17:04,880 Yeah, I really want to reiterate that, you know, Gen A high is all this new stuff and 259 00:17:04,880 --> 00:17:08,160 large language models and blah, blah, blah, blah, blah. 260 00:17:08,160 --> 00:17:10,400 But you can never lose track of the basics, right? 261 00:17:10,400 --> 00:17:14,760 I mean, good secure application development, good secure application design. 262 00:17:14,760 --> 00:17:18,560 You should never ever lose track of the basics, even though it's this new whiz bang feature 263 00:17:18,560 --> 00:17:19,560 thing. 264 00:17:19,560 --> 00:17:22,840 So yeah, I really want to reiterate that. 265 00:17:22,840 --> 00:17:26,100 I also want everyone to know who's listening that, you know, we're really going to focus 266 00:17:26,100 --> 00:17:27,140 on some of the depth. 267 00:17:27,140 --> 00:17:32,200 This is really an episode to sort of set the frame for the next three episodes where we're 268 00:17:32,200 --> 00:17:33,920 going to a lot more depth. 269 00:17:33,920 --> 00:17:38,480 I think it's really important that everyone understands kind of what the problem space is 270 00:17:38,480 --> 00:17:43,360 and how we're sort of working on it and how you, you know, as a consumer of these products 271 00:17:43,360 --> 00:17:47,520 should also understand about how to, you know, how to protect the environment. 272 00:17:47,520 --> 00:17:51,160 By the way, just so if anyone listening is not aware, the prior episode to this, we actually 273 00:17:51,160 --> 00:17:57,800 spoke about some aspects of oversharing using basically blueprints, blueprints with a lower 274 00:17:57,800 --> 00:18:02,120 case B, not as your blueprints with an upper case B, but just blueprints, like ideas about 275 00:18:02,120 --> 00:18:06,520 how you can actually go around securing gen AI applications. 276 00:18:06,520 --> 00:18:09,960 In this one, we're going to go in a lot more detail and some of the other aspects that 277 00:18:09,960 --> 00:18:14,680 are, you know, essentially the basics of securing environments for, you know, gen AI environments. 278 00:18:14,680 --> 00:18:16,120 So I think this is good. 279 00:18:16,120 --> 00:18:17,120 All right. 280 00:18:17,120 --> 00:18:18,480 So we've gone over some of the examples. 281 00:18:18,480 --> 00:18:23,600 I mean, we could, you know, we could talk about, you know, example, you know, sort of 282 00:18:23,600 --> 00:18:28,520 violating gen AI models and getting them to do things they shouldn't do until a blue in 283 00:18:28,520 --> 00:18:31,760 the face as many, many, many examples. 284 00:18:31,760 --> 00:18:37,840 So Diana, why don't you just give a quick overview of each of the other two, like what's 285 00:18:37,840 --> 00:18:41,560 in episode two, what's in episode three, and then what's in episode four? 286 00:18:41,560 --> 00:18:42,560 Yeah. 287 00:18:42,560 --> 00:18:46,720 So as I said, and you were also saying, this is just an intro episode to what's to come 288 00:18:46,720 --> 00:18:49,920 in the next, the next couple of episodes. 289 00:18:49,920 --> 00:18:55,280 So my colleagues from the identity space, as I said, we wrote this MSLearn article, 290 00:18:55,280 --> 00:18:59,360 which is all about how to secure gen AI applications with Entra. 291 00:18:59,360 --> 00:19:04,040 So in the next three episodes, we're going to dive deep into the specifics and more technical 292 00:19:04,040 --> 00:19:06,440 aspects of using these Entra capabilities. 293 00:19:06,440 --> 00:19:11,680 So for example, in episode two of this series, my colleague is going to talk about how to 294 00:19:11,680 --> 00:19:14,480 discover over permission issues with Entra. 295 00:19:14,480 --> 00:19:19,480 So how to use Microsoft Entra permissions management to identify, to manage over permission 296 00:19:19,480 --> 00:19:23,560 identities in multi-cloud environments to reduce security base. 297 00:19:23,560 --> 00:19:28,200 So this is very important and they're definitely going to talk about, you know, other real 298 00:19:28,200 --> 00:19:33,520 life examples as well, but getting into the deep technical aspects of it is going to be 299 00:19:33,520 --> 00:19:34,520 so helpful. 300 00:19:34,520 --> 00:19:38,680 So for anyone interested in learning more about, please, please check out episode two 301 00:19:38,680 --> 00:19:39,680 of this series. 302 00:19:39,680 --> 00:19:45,400 Then in the next episode, my colleague is going to talk about enabling access control. 303 00:19:45,400 --> 00:19:50,220 So she's going to talk about Microsoft Entra ID governance and how lifecycle workflows 304 00:19:50,220 --> 00:19:54,600 can help you also prevent many of the situations that we discussed today. 305 00:19:54,600 --> 00:19:59,800 And then finally, in the fourth episode of this series, we're going to talk about monitoring 306 00:19:59,800 --> 00:20:00,800 access. 307 00:20:00,800 --> 00:20:07,040 So how you can use Entra permissions management, Entra ID governance to monitor the access, 308 00:20:07,040 --> 00:20:11,720 to make sure that everything is okay in your environment and that you are preventing this, 309 00:20:11,720 --> 00:20:13,640 the situations that are awful, right? 310 00:20:13,640 --> 00:20:19,040 Nobody wants people in their company, in their organizations to have to experience this type 311 00:20:19,040 --> 00:20:21,520 of unintentional AI discovery situation. 312 00:20:21,520 --> 00:20:24,760 So the next three episodes are going to be very interesting. 313 00:20:24,760 --> 00:20:28,920 So if you're, if you're excited to learn more about the technical aspects and these capabilities 314 00:20:28,920 --> 00:20:33,960 that we briefly discussed about today, please, please join the whole series. 315 00:20:33,960 --> 00:20:35,520 Fantastic. 316 00:20:35,520 --> 00:20:40,960 So with the sort of intro episode out the way, one thing we always ask our guests, Diana, 317 00:20:40,960 --> 00:20:46,120 is if you had just one thought to leave our listeners with, what would it be? 318 00:20:46,120 --> 00:20:48,640 Yeah, so that's a, that's a great question. 319 00:20:48,640 --> 00:20:53,680 I, as I said at the beginning of this episode, I think it's really important to learn about 320 00:20:53,680 --> 00:20:55,720 this topic specifically, right? 321 00:20:55,720 --> 00:21:01,160 To learn more about what happens when you are using general applications in your, in 322 00:21:01,160 --> 00:21:04,920 your organization and you implemented them into your operations. 323 00:21:04,920 --> 00:21:10,800 It's important not only, you know, like as work, someone who works as part of our organization, 324 00:21:10,800 --> 00:21:14,480 but even if you are just interested in learning more about AI, I feel like this is a very 325 00:21:14,480 --> 00:21:19,240 important aspect of the application of AI in operation. 326 00:21:19,240 --> 00:21:23,360 So definitely, you know, like take some time to read about it. 327 00:21:23,360 --> 00:21:28,360 There's plenty of information out there to learn more about this topic and please check 328 00:21:28,360 --> 00:21:32,680 out the upcoming episodes on, on, on security engineering with Entra. 329 00:21:32,680 --> 00:21:33,680 Okay. 330 00:21:33,680 --> 00:21:34,960 Let's bring this episode to an end. 331 00:21:34,960 --> 00:21:37,200 Again, Diana, thank you so much for joining us this week. 332 00:21:37,200 --> 00:21:38,880 I've been very, very busy. 333 00:21:38,880 --> 00:21:42,400 And again, just to level set, you know, this is an introduction to the other three, the 334 00:21:42,400 --> 00:21:45,160 other three episodes that are coming out. 335 00:21:45,160 --> 00:21:47,760 We hope to do those in relatively rapid succession. 336 00:21:47,760 --> 00:21:51,000 So to all our listeners out there, we hope you found this intro of use, at least just 337 00:21:51,000 --> 00:21:53,680 give me a background as to what the problem space actually looks like. 338 00:21:53,680 --> 00:21:58,040 And again, we'll, we'll show some real solutions in the next, next three episodes. 339 00:21:58,040 --> 00:21:59,400 So again, thank you for listening. 340 00:21:59,400 --> 00:22:01,200 Stay safe and we'll see you in the next one. 341 00:22:01,200 --> 00:22:04,120 Thanks for listening to the Azure Security Podcast. 342 00:22:04,120 --> 00:22:10,960 You can find show notes and other resources at our website, azsecuritypodcast.net. 343 00:22:10,960 --> 00:22:16,120 If you have any questions, please find us on Twitter at Azure Setpod. 344 00:22:16,120 --> 00:22:44,680 Creative Music is from ccmixtor.com and licensed under the Creative Commons license.