النص 🔗
WEBVTT
00:00:07.040 --> 00:00:08.320
Welcome back everybody.
00:00:08.400 --> 00:00:11.679
My name's Tommy McCarthy and this is Let's Talk Cyber.
00:00:12.000 --> 00:00:17.039
I am delighted to welcome a future superstar of cybersecurity.
00:00:17.280 --> 00:00:22.559
I met this young lady, Afra Al Mansori, some years ago in Dubai at GISEC.
00:00:22.800 --> 00:00:29.920
We met again in Abu Dhabi and I was delighted to catch up with Afra at the Global Government Cloud Forum back in January.
00:00:30.160 --> 00:00:34.000
I invited her to come along and share some of her views and opinions on cybersecurity.
00:00:34.079 --> 00:00:35.520
And we're going to talk about AI.
00:00:35.600 --> 00:00:38.719
But first of all, Afra, welcome to Let's Talk Cyber.
00:00:38.960 --> 00:00:39.520
Thank you, Tommy.
00:00:39.840 --> 00:00:41.039
It's a pleasure to have you here.
00:00:41.280 --> 00:00:42.159
Remember what I said.
00:00:42.240 --> 00:00:44.799
We're here to share some knowledge with our with our viewers.
00:00:45.119 --> 00:01:00.479
Afra, we're going to talk a little bit about AI, but before we do, one thing that's clearly evident to not just me, but everybody else that's out there in the cyber community, realizing AI is taking over the world at a rapid, rapid rate.
00:01:00.719 --> 00:01:11.359
And it's a it's a delight that we've got young professionals like you, an avid member of the women in cybersecurity in the Middle East, being championed by the leaders there in that group.
00:01:11.519 --> 00:01:15.280
And hopefully one day we'll see you there at the top of the ladder.
00:01:16.159 --> 00:01:29.920
It's clear to me that cybersecurity needs to keep up with the trends and keep up with the times, and AI is certainly something that we all need to take a real hard grip of because it it's it's expanding on such a rapid pace.
00:01:30.159 --> 00:01:35.040
What I want to do is ask you a few questions about your views and your opinions about AI.
00:01:35.439 --> 00:01:40.560
What's one of the things about AI and cybersecurity that most people completely misunderstand?
00:01:40.719 --> 00:01:43.280
What are people not getting about AI and cybersecurity?
00:01:43.760 --> 00:01:45.840
It's a good question, uh Tommy.
00:01:45.920 --> 00:01:57.280
Uh many believe that AI is like a magic bullet, like a fully autonomous solution that once deployed, it will automatic uh automatically secure systems without any human involvement.
00:01:57.519 --> 00:02:07.760
But in reality, AI in uh cybersecurity is a powerful tool designed to augment human expertise rather than actually replacing them.
00:02:08.000 --> 00:02:13.360
So here the misconception lies in uh overestimating the capabilities of AI.
00:02:13.439 --> 00:02:29.280
And while it can process and analyze vast amounts of data far faster than uh human expertise, it still heavily depends on the quality of training, the program, the algorithms even in use, and the continuous human oversight.
00:02:29.520 --> 00:02:39.680
So uh for example, uh, as you know, AI systems may struggle with novel or adaptive threats unless they are regularly updated with the fresh intelligence and data.
00:02:39.840 --> 00:03:03.680
And uh moreover, uh adversaries are now employing AI and uh AI techniques themselves, creating a uh arms race where uh contextual uh misunderstanding or understanding ethical considerations and even strategic um judgment uh are crucial uh areas where human insights, like from my point of view, uh remains undeceasable.
00:03:04.560 --> 00:03:06.800
Yeah, you just touched on something really interesting there.
00:03:06.879 --> 00:03:09.919
And I think this is what people are misconceiving about AI.
00:03:10.080 --> 00:03:14.479
There's still a human interaction when it comes to expanding the capability of AI.
00:03:14.560 --> 00:03:16.319
Would I would I would I be right in saying that?
00:03:16.479 --> 00:03:17.439
Is that your humble opinion?
00:03:17.759 --> 00:03:21.360
Yes, from my point of view, it still requires uh human intervention.
00:03:21.520 --> 00:03:36.639
Like uh of course, uh the algorithms when they are going to be used, we have to have actual data, actual uh data that can be used in uh detecting, for example, anomalies or uh understanding how cyber threats uh are uh occurring.
00:03:37.199 --> 00:03:37.759
Yeah, good.
00:03:37.840 --> 00:03:40.960
Well uh you know then it's interesting because it leads me on to my next question.
00:03:41.280 --> 00:03:51.520
So how can AI look to if well how can AI transform cybercu cybersecurity from a from a uh a reactive to more of a proactive approach?
00:03:51.680 --> 00:03:52.800
What are your views on that?
00:03:53.360 --> 00:04:01.439
Um You know, uh Tommy, traditional uh cybersecurity is uh like relying on predefined rules currently.
00:04:01.599 --> 00:04:09.199
So making it reactive and often uh can often really slow the uh encounter of new threats, from my point of view.
00:04:09.439 --> 00:04:28.879
So as AI is used, it AI actually shifts this approach by continuously analyzing uh vast amounts of data, as I specified before, and uh it really detects anomalies uh and predicts the threats that uh might happen or might happen before uh the incident might occur or the threat might occur.
00:04:29.040 --> 00:04:38.160
So machine learning enables AI to uh identify these patterns, automate uh threat hunting, and even adapt defenses in uh real time.
00:04:38.639 --> 00:04:50.000
So uh there is a risk, however, uh when AI is introduced, as AI uh also introduces uh risks as adversarial uh risks, as I specified before.
00:04:50.240 --> 00:04:57.519
Adversarial uh attacks uh manipulate AI models really as we are moving on to the uh new technology.
00:04:57.680 --> 00:05:08.560
So to ensure the effectiveness actually of um the organizations, for example, we uh organizations must secure AI systems through uh adversarial training.
00:05:08.800 --> 00:05:15.120
This is one point, and we have to
Load full transcript
explain how or train how AI techniques are uh are used.
00:05:15.279 --> 00:05:27.279
So while AI uh introduces uh and enhances cybersecurity, human expertise remains a crucial uh investment in overseeing and improving the decision making.
00:05:27.439 --> 00:05:42.319
But by integrating uh AI uh with traditional uh security measures, uh organizations from my point of view can see ahead uh of cyber threats that might occur and build a more uh resilient uh cyber defense system.
00:05:43.120 --> 00:05:45.519
Well, it's it's that's quite interesting, actually.
00:05:45.680 --> 00:05:48.000
Did I ask did I ask the right question there?
00:05:48.079 --> 00:05:50.399
Because you certainly gave me enough of a response.
00:05:50.720 --> 00:05:53.600
Let me ask you something that's not in our question bank.
00:05:54.240 --> 00:05:58.399
Are you using AI in your everyday life, either for work or for pleasure?
00:05:58.720 --> 00:05:59.279
Of course.
00:05:59.360 --> 00:06:04.480
Uh currently, uh like for example, in my daily life, I use AI for everything.
00:06:04.720 --> 00:06:16.319
I use uh AI for um uh completing tasks, uh AI for implementing a new code, for example, for a website, or uh other like uh tasks that might uh be included in my life.
00:06:16.480 --> 00:06:23.680
So AI is a core, uh from my point of view, it became a core uh aspect in um our daily lives and like using it.
00:06:24.399 --> 00:06:29.279
So how do you know if AI is not influencing this the the decisions that you make?
00:06:29.519 --> 00:06:38.000
How do you know that you're making the the decisions for how you approach projects on your own basis as opposed to allowing AI to influence you?
00:06:38.160 --> 00:06:38.959
How do you manage that?
00:06:39.519 --> 00:07:00.399
For example, like uh from my uh from my perspective, I use AI to enhance, like, for example, my thought or give me more knowledge of how of how like, for example, I can implement stuff either uh even if they are like code-based or uh uh cybersecurity, for example, uh new projects or uh tools that might be used.
00:07:00.480 --> 00:07:05.439
I don't use AI to fully uh automate or uh conduct my tasks.
00:07:05.600 --> 00:07:07.759
I use AI to enhance how I work.
00:07:08.079 --> 00:07:10.240
Apologies for throwing that question in there, Afro.
00:07:10.480 --> 00:07:12.160
But let me tell you that was an awesome response.
00:07:12.240 --> 00:07:22.720
So it clearly uh it clearly defines to me that you you understand how to manage your AI interactions and and how to look at the benefits that AI is bringing to your day-to-day work life.
00:07:22.959 --> 00:07:24.639
Now then, the bad guys.
00:07:24.879 --> 00:07:26.480
Yes, the cyber criminals.
00:07:26.720 --> 00:07:32.480
They're clearly using AI to leverage their ability, um, you know, obviously in cyber attacks.
00:07:32.639 --> 00:07:36.560
How can we um defend against these uh bad guys using AI?
00:07:36.639 --> 00:07:39.040
What we what are we doing to combat the bad guys?
00:07:39.439 --> 00:07:43.680
Uh uh there are many points that like I can address in uh this question.
00:07:43.920 --> 00:07:52.000
Cyber criminals are really using AR or leveraging AI to launch advanced, automated, and highly targeted attacks.
00:07:52.240 --> 00:08:01.040
As you know, we're moving into an era of using a lot of new technology tools that might be implemented or used uh in a good way or in a bad way.
00:08:01.279 --> 00:08:10.879
So AI-powered phishing creates really realistic uh emails or texts, and deepfake, as you know, deepfake has become a trend in uh the past uh months.
00:08:11.040 --> 00:08:14.800
So increasing the success of uh social engineering.
00:08:15.040 --> 00:08:26.959
So adaptive malware and uh ransomware really use AI to evade detection and modifying their behavior in real time to bypass codes or bypass security measures.
00:08:27.199 --> 00:08:34.240
So here we're talking about a good uh infrastructure or a good uh uh architecture of uh an organization.
00:08:34.480 --> 00:08:40.639
So uh security professionals will bypass this in updated uh security measures have been implemented.
00:08:40.879 --> 00:09:10.480
So uh to encounter, from my point of view, uh these threats, um, there are many points that can be addressed, like organizers organizations must uh implement AI-driven cyber security solutions uh capable to um adapt or uh uh detect real-time threat detection uh and anomaly analysis, which is a key point here, and uh the adversarial training to strengthen uh strengthen these defenses.
00:09:10.720 --> 00:09:33.600
So um here we will ensure the transparency and security decisions and uh while continuous updates are occurring uh uh by human oversight or by the AI uh tools, uh human oversight, uh like I have to point uh this note is a critical in identifying the complex threats that might occur.
00:09:33.919 --> 00:09:38.399
So we are not fully dependent on AI from my point of view.
00:09:38.559 --> 00:09:55.279
And um a combination of AI and strong governance and proactive um threat intelligence sharing uh is essential in staying ahead of uh cybersecurity threats that might uh we that we might encounter in the future.
00:09:56.000 --> 00:10:04.720
Well, certainly governance and I believe ethical use of AI will have to be a part of every day's uh everybody's day-to-day interactions.
00:10:04.879 --> 00:10:13.039
When you're looking at using AI tools, I think you have to be you have to be guided by the governance obligations you have as a professional.
00:10:13.200 --> 00:10:17.840
And then ethically, I think it's really important ethically that you do the right thing.
00:10:18.159 --> 00:10:26.720
How do you stay innovative um and and creative when it comes to you know um keeping up with AI and cybersecurity?
00:10:26.799 --> 00:10:29.360
How how are you gonna stay uh um sort of creative?
00:10:29.759 --> 00:10:32.639
Um I say uh innovative in many aspects.
00:10:32.720 --> 00:10:37.679
Like I say innovative in the field um field by continuous learning.
00:10:37.919 --> 00:10:54.720
And I have a I allocate really time in my um my time in reading the latest research papers or even industry reports or new news might be um critical in uh AI, for example, or cybersecurity, or the latest trend in uh technology.
00:10:54.879 --> 00:11:00.000
And I have engagement really with outside uh traditional like tech boundaries.
00:11:00.159 --> 00:11:15.279
I really uh network with good industry leaders in the cybersecurity field to be able uh to get insight from them as they are uh uh they have implemented or they have used these tools in uh their organizations.
00:11:15.600 --> 00:11:27.360
And um as as this is a uh point that might be important in being innovating and staying innovative, of of course, experimentation is a um uh critical aspect here.
00:11:27.519 --> 00:11:33.519
If you don't uh uh experiment tools in AI or cybersecurity, you won't be able to actually defend yourself.
00:11:33.679 --> 00:11:42.399
You have to be uh staying updated uh on the the latest tools and trends, and even if just conducting it in um like a sandbox environment.
00:11:42.799 --> 00:11:49.360
This is really practical from my point of view as I enter the cybersecurity fleet field in the young age.
00:11:50.080 --> 00:11:53.120
Well, then I guess after wow, what an amount of information.
00:11:53.279 --> 00:11:55.600
Are you sure you're uh are you the real deal?
00:11:55.759 --> 00:11:59.200
Are you human or are you an AI autobot that I'm speaking to right now?
00:11:59.440 --> 00:12:06.159
Because that is an unbelievable amount of information about AI, um, but we have a secret for you as well.
00:12:06.399 --> 00:12:17.919
I'm gonna ask you if you had to sum up your own views on AI and cybersecurity philosophy, what is the one sentence that you'd give me to sum up AI philosophy as you see it?
00:12:18.559 --> 00:12:23.200
This is a very good question that is really hard for me to uh answer.
00:12:23.279 --> 00:12:27.120
But if I'm gonna be a long it's gonna be a long sentence, I can tell right now.
00:12:27.440 --> 00:12:30.480
It's going to be a one sentence to wrap up everything.
00:12:30.720 --> 00:12:44.559
So uh AI for me is a powerful tool, Ali, that uh really uh will be a key uh point in the cybersecurity industry, but its true strength is lies in a balanced approach.
00:12:44.720 --> 00:12:53.840
So combining uh automation, adaptability, and human intelligence will be a key point to stay updated on the evolving threat that might occur.
00:12:54.000 --> 00:12:55.919
So we are playing a two role.
00:12:56.080 --> 00:12:58.159
It's AI versus human.
00:12:58.879 --> 00:13:01.039
So now, was that in one breath?
00:13:01.279 --> 00:13:03.600
Did you have to breathe in in that sentence?
00:13:03.759 --> 00:13:06.720
Because if you if you didn't have to breathe, we'll class that as a sentence.
00:13:06.799 --> 00:13:07.360
Well done.
00:13:07.519 --> 00:13:22.799
Look, Afra, that was that's an awesome insight from a young person's perspective as to how you see AI adding value, how you see AI becoming the game changer in the world that you now live and work in, cybersecurity field.
00:13:23.039 --> 00:13:33.679
As you know, we're working towards the GISEC um conference in less than a month's time, and there's gonna be so much going on, including a generative AI workshop by OSP Cyber Academy.
00:13:33.840 --> 00:13:36.960
So we would expect to see you and and your gang there.
00:13:37.679 --> 00:13:40.480
I know for a I know for a fact, there's no question.
00:13:40.639 --> 00:13:47.200
What I see in front of me is a future superstar of cybersecurity, the next and the youngest CISA in Dubai.
00:13:47.279 --> 00:13:53.279
I have no doubt that you're gonna raise to that level, Afra, because your passion is unquestionable.
00:13:53.440 --> 00:13:57.279
Everywhere there's anything to do with cyber, you always seem to be there.
00:13:57.360 --> 00:13:59.120
How on earth you find the time, I don't know.
00:13:59.360 --> 00:14:03.120
Afra al Mansori, the superstar of cyber security.
00:14:03.360 --> 00:14:03.519
Thank you.
00:14:03.919 --> 00:14:06.559
Thank you so much for so much for joining me on Let's Talk Cyber.
00:14:06.720 --> 00:14:07.600
Thank you so much, Tommy.
00:14:07.679 --> 00:14:08.879
It was a privilege.
00:14:09.279 --> 00:14:10.159
Wonderful.