• 3 years ago
“If you are using AI, you need to defend new stuff that you may have not defended before.”


Check Point CTO Dr. Dorit Dor discusses the evolution of AI in cybersecurity in a conversation with Niraj Shah at #Davos2024. #NDTVProfitAtWEF 





Category

📺
TV
Transcript
00:00 with cyber security as well, both on enabling cyber security to defend oneself,
00:05 whether businesses or governments, and to the use of AI for ransomware or other stuff.
00:12 Now, how will this evolve? And do you reckon that as a percentage of the P&L,
00:19 the quantum of costs that a company or a CIO of a company has to bear for the next few years,
00:26 the cost of cyber security will only inch up because of the sophistication of the attacks
00:31 that we've seen recently? So these are a collection of questions,
00:35 I'll do my best to answer all of them. So first about the impact of AI. The first immediate impact
00:42 is more targeting and more automation. So look at everything you did before, you could now do it in
00:50 a higher scale and more automated scale. You can, in a higher scale and more automated way,
00:56 attack somebody, and you could, at a higher scale and more automated way, defend somebody.
01:01 So this will increase the sophistication, and when there is an increase in sophistication of
01:07 an attack, there is an increase in cost associated with it. So there is an increase of this.
01:14 But there is also another layer of everybody's using AI now, and if you are using AI,
01:22 you need to defend new stuff that you may have not defended before. So now as an organization,
01:27 you collect more data. And now you have to worry about who has access to your data internally.
01:33 Before, only seldom people had access and only to seldom parts. Now you want all the data to
01:39 coherently live in your data lake, and you want to give it access to all the AIs around you that
01:46 activate it. So now you have to guard your jewelry much more and secure how you do your own AI. So
01:56 there is the attacker AI, there is the defender AI, and there is how you defend your AI usage,
02:02 what comes out of it, how you defend the models, how you defend what AI you are using, and how you
02:08 defend your data. These are all three things. The other thing I would say is usually there are two
02:15 elements in technology transformation. A, we always think it would run faster than it actually
02:21 is, so things do take time, so it's not all changing in a day. But B, looking at different
02:28 changes that happened in the past, we could see that it starts from these obvious things. So yes,
02:35 everything will be automated, everything will be conversational, you'll get your answers much more
02:39 clear to a bigger audience. Sixth level could get the answers as much as a technician. But at the
02:45 same time, there would be an evolution of idea of things that would be done differently, that we
02:52 don't know to believe now how they would be done differently. But some of the things would not be
02:58 just automated the way it's done today, but a different way to do them given the tools. And
03:05 that will change things. If you look at the mobile world, we do today quite a few things very
03:10 differently than we did at the beginning of the mobile. We anticipated talks and maybe some other
03:17 in text and maybe picture moving, but we are doing today much more applications and much more
03:23 personalized with the mobile. So the same change will happen with AI. The AI will start from the
03:30 obvious automation and conversation and ease of use, etc., but would move to create new use cases
03:37 and new ways to conduct our business. Maybe some things would be direct to consumer because the
03:43 consumer would use the AI as guidance to a company. Maybe some things will be passing the humans or
03:53 in the loop in some sense. I don't know to tell you because one of the things is it's hard to
03:58 predict. But suffice to say that if increased users of cloud, so availability of data everywhere, AI
04:06 plus sophisticated ransomware would mean that the cost of an average organization or a government
04:13 for defending itself against all of these would mean increased spends on cybersecurity.
04:20 Yes, but I also want to say that they have increased spends on the underlying technology.
04:25 So now they spend also on the AI and more on software. In general, we see an increase
04:31 of spending in software in the recent years. If you look at production line, industrial line,
04:37 they have much more software components. You look at things that were in hardware, a lot of it is
04:42 kind of software, firmware changes. So you have to look at the expenses of security also in light of
04:48 how much you are spending on these elements. So the spending changes and with it there is
04:54 more spending on cybersecurity, but there is also more spending on the cyber usage.
04:59 [Music]
05:09 [BLANK_AUDIO]

Recommended