- 4 mesi fa
Categoria
📺
TVTrascrizione
00:05Benvenuti e ben ritornati a questo speciale dedicato allo stretto legame tra AI e Cyber Security,
00:11la cosiddetta Cyber Wake Up Call, quella che è stata definita negli ultimi mesi dopo le sfide
00:17esposte dal mitos, il modello di Antropic, che ha dimostrato vulnerabilità al mondo software che
00:25finora erano sconosciute, questo ha allarmato governo in banche centrali e istituzioni. Per parlare di
00:30come l'AI stia rivoluzionando anche il settore della Cyber Security nel contesto globale, ne parliamo con
00:37un grande esperto, Amir Lakhani, Senior AI Director AI at Research di Fortinet, è un grande piacere.
00:47Allora Amir, partiamo dalla prima domanda, cioè che come sta cambiando l'AI in Cybercrime? L'abbiamo
00:53detto ci sono il tema degli agenti autonomi, mitos, come si attacca e come ci si difende?
01:01No, that's a great question. I think first of all, when we talk about AI and cybercrime,
01:08there's a lot of Hollywood-ness that comes out in this and we have to keep it real what
01:13really is out there. But AI is changing the game for both the good guys and for the bad guys.
01:20For
01:20the bad guys, ever since I started cyber security, what the bad guys did is they waited for a
01:25problem, a vulnerability that came out and they wrote an attack against that vulnerability.
01:30AI has changed the game. Now AI knows about all the vulnerabilities, it just waits for them
01:36to be released and starts writing code, starts writing code to attack systems. It makes it
01:42much easier, but not only easier, much faster for the bad guys to attack. But I'm pretty excited,
01:48as a cyber security professional, I think we can now use AI. The first time in my career
01:52I feel like we have a winning chance, at least a chance against the bad guys because as fast
01:59as AI makes it easier for the bad guys, it also gives us that same capabilities as good
02:05guys to defend. It's still a cat and mouse game and of course the bad guys, they need very
02:11little to succeed while as good guys we need to look at a lot of things to succeed. So it's
02:16still
02:16dangerous, but AI is making it easier for the bad guys to do more things and to go faster.
02:22Quindi sta diventando più facile per gli attaccanti e allo stesso tempo più
02:27difficile per chi deve proteggere. Ma venendo invece all'Italia, come si posiziona l'Italia
02:34in questo contesto? Qual è il target di fatto d'Italia e come si contestualizza in confronto
02:40in questo contesto?
02:43No, Italy, that's a great question, because Italy is in a very interesting position because
02:47you have many different industries, industries in energy, in operational security, in critical
02:54infrastructure, as well as a lot of businesses. Because of that, this is a very, very rich
03:00area for the bad guys. They have a lot of places that they can attack. And there's a feeling
03:06that in Italy that if they attack these places, if they do ransomware attacks, that they'll
03:12be more successful in getting paid.
03:15Now, one of the things that's interesting in Italy that we saw from our stats is that a
03:20lot of the bad guys are getting into companies by not actually attacking them, not using new
03:26attacks, but stealing passwords. And I thought that was very interesting because, you know,
03:30there's a lot of ways we can protect against passwords. But what we're finding is people just
03:36aren't using good common security sense. They're not using best practices. Hackers are not
03:42hacking into companies in Italy. They're just logging in. They're sending software to steal
03:47your passwords. They're sending phishing emails or fake emails or fake text messages that
03:52people are clicking on, stealing the passwords, and then using that to get into systems. Of course,
03:58the nice thing about AI is we can detect when these vulnerabilities and these risks exist, be proactive
04:04about them and start making some changes to mitigate and stop these attacks.
04:08And then, Amir, starting from Italy, Italy can be a good model to understand how the
04:15scenario of cyber security is evolving and threats. But on the global scale, how do you see
04:20this evolution and what happened in the last few years? We also think about the conflicts in
04:26the world. Cyber security is always more an element of attack in these contexts. What are the
04:34minacce that you care about?
04:37So, as Italy becomes more prominent on the world stage, they're becoming more of a leader
04:43in our world affairs. Because of that, they're gathering the attention of what's happening in
04:48the Middle East. A lot of people have a lot of opinions. A lot of people within the Middle East
04:53have a lot of opinions on Italy's leadership. So they're gathering a lot of attention from the
04:59Middle East, from places like China, from North Korea. And because of that, the world view, Italy
05:07is kind of making themselves a target. But they have to do that because they should be a leader.
05:12They should understand what's happening in cyber security. And they need to be able to defend
05:17against the possible threats. So I think this will happen with any country. Any country that takes
05:22any type of stance, you're going to have people that agree with it, people that don't agree with it. And
05:27that's
05:27going to cause some tension and the tension. And sometimes you have the bad guys targeting you
05:33because of your views and because of your stance and your opinions.
05:37Ecco, stavi dicendo giustamente che con l'intelligenza artificiale appunto il lavoro un po' diventa
05:44più complicato dal nostro punto di vista. Quindi a netto di tutto quello che hai raccontato finora, quali
05:49sono le strategie difensive che effettivamente bisogna mettere in campo per rispondere
05:54e apposite relazioni.
05:55Now, that's also a good question because we see so many people, so many companies implement
06:01AI, get on what we call the AI bandwagon. But a lot of times they're not thinking about
06:06what happens if something goes wrong. What are my backup solutions for AI? What happens
06:12if AI doesn't work? And how can AI be broken? And that's the most important thing is we recommend
06:18to our clients, to our customers, test your AI systems like you would any other critical
06:23system, test what the vulnerabilities are, test how people can exploit those and get into
06:28those systems. Because there are, today AI is how computer systems were 20 years ago.
06:34There could be new problems that are found every day and every day you have to protect
06:38against those problems. So our solution really is to test your systems, understand the risks
06:45risks and close the risks. Bad guys are always going to attack you. And sometimes they're
06:50really smart. Sometimes they can be smarter than us. So we gotta, we gotta plan for that.
06:54We gotta plan for multiple levels of protection.
06:57E una cosa aggiungo, Amir, ovviamente bisogna investire in cyber security e proteggersi perché
07:03in Italia purtroppo il contesto è ancora molto esposto e la cyber security non è ancora una top
07:10priority. Nonostante, citando solo il World Economic Forum, la cyber security sia una delle maggiori
07:16preoccupazioni per il CEO. So, thank you again, Amir. Thank you for being with us.
07:22It's a great pleasure and I hope to speak to you soon again.
07:24Hey, thank you very much.
07:26E grazie per Amir, la canica di essere stato con noi.
07:28E' stato un grande piacere averla per questo speciale. Ci rivedremo in serie.
07:34Grazie per essere stati con noi e ci rivediamo sempre qui su Class. Servici.
07:38Grazie.