Dr. Craig from Cohesity joins us to explore the role in boardroom decisions, and how it’s shaping the future of cybersecurity. From defending against ransomware to unlocking the power of future-ready data, we’ll unpack the strategies that can help Malaysian organisations stay resilient and competitive in today’s high-stakes digital landscape.
Category
🗞
NewsTranscript
00:00Selamat datang kembali ke Niagawani.
00:02Di seluruh dunia, kita melihat kemampuan penyakit perangai atau perlengkapan data di sektor industri berbeza.
00:09Dari DAVITA, sebuah firma yang berlaku penyakit perangai yang terkenal,
00:12sebuah penyakit perangai yang terkenal di Toppenax Tech atau Team NS di Singapura,
00:18untuk penyakit perangai yang meminta 10 juta USD di perangai Malaysia,
00:23di perangai yang berlaku di perangai yang berlaku di perangai yang berlaku di Malaysia.
00:26Dan ini menyebabkan perangai yang terkenal untuk perangai dan perangai
00:31untuk mengembangkan kemampuan penyakit perangai mereka semasa penyakit perangai
00:35kemampuan penyakit perangai lebih agresif dalam kemampuan dan kemampuan mereka.
00:39Ia penting untuk penyakit perangai industri untuk mengembangkan data proaktif,
00:43mengembangkan infrastruktur digital dan mengembangkan strategi penyakit perangai yang berlaku.
00:50Dan bersama saya sekarang, Dr Craig Martel,
00:54sebagai Chief AI Officer of Cohesity.
00:56First of all, thank you so much for coming to our show today.
01:00Maybe, what is the state of artificial intelligence AI today?
01:05So first of all, terima kasih.
01:07Thank you for having me.
01:09So, AI today is a really interesting spot in the hype cycle.
01:14We had this really large peak in all of the magic that AI was going to do,
01:20and now we're starting to come down from that hype cycle
01:23and see where it's really effective and where it's not.
01:26So, I think AI is showing really strong promise for things like helping with coding,
01:33helping with writing in the sciences.
01:37But I think it's really important to think about AI not as a singular technology,
01:43such that if you buy it, you can solve all of your problems,
01:46which is the way it's sold.
01:48Instead, you need to think it's a set up, it's a collection of technologies,
01:52and it may work for one use case, but may not work for another use case.
01:56Cybersecurity and AI are top of mind or bot rooms,
02:01and what is AI's critical role in safeguarding data against ransomware and other threats?
02:08So, there's a number of ways to approach that.
02:11The first thing I would say is that the same power that allows attackers using AI
02:19to get easier access to your data also allows you to protect it, right?
02:24So, for example, you could use AI to red team.
02:27Try to find those ways into your data themselves.
02:31But other things that are really important is in your backup system
02:35to automatically scan for malware.
02:37Can you detect whether there's something there that shouldn't be there?
02:40And most importantly today is can you detect a change in entropy?
02:45Because entropy is an indication of a ransomware attack.
02:48And you can use AI to do these things.
02:50Oh, okay.
02:51Okay.
02:52And then what data protection and management strategies will help organizations,
02:57especially if you see in Malaysia, focus on driving positive business outcomes?
03:03So, let's just go back to that ransomware notion for example.
03:06When you think about data protection, you really should think about the NIST,
03:10the US National Institute of Standards, notion of 3-2-1.
03:15three copies in two locations, one of which is immutable, not changeable.
03:20So, you should think about your, and ideally that copy is air-gapped.
03:24So, completely separate from the rest of your company's data.
03:28Because if you detect that there's a ransomware attack,
03:32you can always go back to the last known good if that data is not touchable.
03:36And so, it's really important when thinking about your data resiliency strategy to follow these sorts of norms.
03:42Yeah.
03:43Maybe how will future ready, especially data, help enterprises have a winning advantage in uncertain times?
03:53So, the really important thing to think about AI in general is that it's, let's ask the question, what is AI?
04:04AI is really statistics at scale.
04:07Okay.
04:08You gather data from the past to build a model to predict the future.
04:12That's the really important thing to think about.
04:14So, given that is statistics at scale, the data that's in a generic AI model may or may not be effective for your use case.
04:24Yeah.
04:25So, the data that's in your backup, the data that you have in your company is the data that's going to provide those advantages.
04:31Okay.
04:32So, any insight into your business, any capabilities that you want to build, that's going to come from the data that's in your business.
04:37Not from stuff external, like if you go to one of the cloud providers and buy AI, that's great.
04:42But that's generic.
04:43Yeah.
04:44The real advantage your company is going to have is leveraging its business data for AI.
04:48Yeah.
04:49Talking about evolving threats, how have cyber threats actually evolved with the rise of AI?
04:56When we are talking about, especially for example, like, are attackers also using AI tools, especially AI tools to scale or automate their attacks?
05:06Yeah.
05:07So, they have been scaling and automating at increasingly big speed, but again, the sort of security folks that we work with have also been doing the same thing simultaneously.
05:19Okay.
05:20So, it is absolutely going to become an arms race, right?
05:22You want to keep up on this arms race.
05:24You want to make sure that you're using the same tools that the bad guys might be using to see if you can find vulnerabilities.
05:31And if I could give a few pieces of advice, there's sort of three things that we at Cohesity recommend.
05:38First, always keep your software up to date.
05:41Because if there's an attack, up to date software may close that attack.
05:45Yeah.
05:46Always use multi-factor authentication everywhere.
05:49A lot of attacks we see attack the identity system of your organization.
05:54So, the bad guys could hijack an identity and so be part of your company in ways you don't want them to be.
06:00But if you use multi-factor authentication, then that's less likely.
06:04And finally, always red team.
06:07Always practice, practice, practice.
06:09Practice what would happen if you were attacked.
06:12Because let's say the worst case happens, you are attacked.
06:15Well, if you've practiced, you can get up and running more quickly.
06:18You can get back to where you need to be more quickly.
06:20Yeah.
06:21So, people tend to think, they tend to think about what should we do after an attack?
06:25Yeah.
06:26Right after an attack.
06:27That maybe is not the best time to think about it.
06:29Yeah.
06:30You should think about that ahead of time, have a plan and practice.
06:32More proactive.
06:33I would like to bring our discussion into data classification and also especially on sensitivity, for example, and zero trust and AI.
06:45Because of data resilience and governance is always interconnected each other.
06:52Sure.
06:53One key challenge is knowing what data to protect most.
06:56and how can actually AI help with data classification and also tagging sensitivity and automating data protection policies.
07:06A hundred percent.
07:07If you think about what, I come from a backup perspective, so I'm going to answer it with respect to, that's what our company does, right?
07:13So, that's what Cohesity does.
07:15So, if you think about that backup data, and if you think about it just sitting in the corner, then you're wasting a very valuable resource.
07:22Because as that data is being backed up, it's representative of your whole business, not just 10 years ago, but also 15 minutes ago.
07:30So, you can use that data as an indicator of there's something being wrong or as an indicator of what data needs to be protected.
07:39So, we constantly run classification techniques where we discover PII, where we discover health information, where we discover other sensitive information that you may or may not want, that you have to govern differently, right?
07:54In addition, we can also do not just malware scanning, but we can look for anomalous behaviors.
08:00So, we start to see, over in this group, more things are being encrypted than were encrypted before.
08:06And it has a lot of PII.
08:08Hmm, maybe that's a ransomware attack.
08:10How about, okay, large enterprises, large companies may have the resources, you know?
08:19But how can small or mid-size, especially in the context of Malaysian businesses, begin their journey towards AI-powered cyber resilience?
08:30You know, maybe your advice on SMEs, how they can start to make sure that their business are well-equipped, is ready on this, yeah.
08:42That's great. That's a great question, and it's something that we at Cohesity think about quite a bit.
08:46Because if you're a large organization, then you already know how to build AI, you already have the tools to build AI, and you already have the problem set that you're concerned with.
08:56We have been thinking hard about how can we offer to our customers sort of an out-of-the-box AI solution that leverages their data.
09:05So, we have a product called Gaia, and Gaia stands for Generative AI Application.
09:10Okay.
09:11It also is the name of a Greek goddess, so we got both out of it.
09:14But Gaia allows you to, just by buying, you know, your backup data, regardless of the size of your business, you have to back up, right?
09:24So, as it's backed up, using our Gaia tool, you can actually ask questions of your data.
09:29Like, you can ask things like, when was the last time our profit looked like the following?
09:36Or, can you show me an example of medical data that may have left illegally or incorrectly?
09:44Yeah.
09:45And it will search through the data, and it will find that, and it will help you answer lots of analytical questions like that.
09:51Yeah.
09:52What role do you think, especially AI, will play in building more cyber-resilient Malaysia?
09:59You know, like, for example, like, and how important public-private partnership, collaboration in this space?
10:07Yeah.
10:08So, I think that, let's address the public-private one.
10:10I think that's really interesting.
10:11Having the right governance rules in place for how you protect your data, how you manage your data, is going to be unbelievably important.
10:20Because, look, you pick up your phone, and you press a button, and suddenly you shared your data with everybody, right?
10:27Yep.
10:28And if it's the Wild West out there, then you don't know where your data is going.
10:31But having a strong governance policy, like I know Malaysia just passed a lot of new AI rules.
10:37Yes, correct.
10:38Having a strong governance policy allows for a uniform set of protections for your data.
10:46So, I think that public-private is really important.
10:49Private doing it alone is great for experimenting and trying new things, but public has to say,
10:56look, this is the minimum set of protections that we need to make sure that our citizens' data is protected.
11:02One of the issues in Malaysia is about the expertise, the technical expertise, especially on AI.
11:09As AI transforms, how can we secure and manage data, and what skills should Malaysian tech leaders and team be developing today?
11:20This is really interesting, given the last talk, the last person you had on.
11:25I think it's really important that education starts to not ignore AI, but integrate AI into the training.
11:34I'm a computer scientist by training, so I know coding best.
11:38Let's talk about coding.
11:40Before AI, you learn a bunch of coding and the programmer codes all of the code and checks for bugs and so on.
11:47Now, the tools are so good that you might code some of it, but there's going to be large chunks of it that you're going to code via AI.
11:55And there's going to be large pieces of it where you're going to ask AI, are there any bugs here?
12:00The problem is, since AI is statistics at scale, as I said before, it's guaranteed to be sometimes wrong.
12:08So it's really going to help people who understand their job be more efficient, more than it's going to replace people with a machine.
12:17So as a coder, you still need the expertise and the understanding to be able to decide that that code generated by the AI was acceptable,
12:25or that it actually did find a bug.
12:28It may not have been a bug, it might have been a design choice on your part.
12:31And so you still need the expertise to understand code, but that coder is going to become significantly more efficient.
12:36But to do that, I think that training with AI, sort of joint coding, human and AI, has to start back in university.
12:43Dr. Craig, it's such a pleasure to talk with you on this matter.
12:48And of course, we, in Malaysia especially, need to make sure that we are resilient enough,
12:56in order to make sure that we have good experts and also our education system are ready to make sure that we are ready to face all the challenges,
13:05and also make sure that AI are integrated together with our education system and make Malaysians more resilient.
13:15Thank you so much, Dr. Craig Martel, Chief AI Officer of Khe Siti.
13:19All of this will be available in our social media platform and astroawani.com.
13:24Thank you so much for watching Astro Awani.
13:27Dan selepas ini, anda akan bersama dengan Tamina untuk meneliti berkenaan dengan satu laporan berkait dengan hidupan liar,
13:36isu gajah yang masih lagi mungkin mendominasi media sosial pada ketika ini.
13:41Kita nak berhenti rehat seketika.
13:42Selepas ini, anda akan bersama dengan Tamina Kaosji dalam Niaga Spotlight.
13:46Terima kasih kerana menonton!