Exploring Network Security and AI Threats with Crystal Redmann

Exploring Network Security and AI Threats with Crystal Redmann

Released Friday, 23rd August 2024
Good episode? Give it some love!
Exploring Network Security and AI Threats with Crystal Redmann

Exploring Network Security and AI Threats with Crystal Redmann

Exploring Network Security and AI Threats with Crystal Redmann

Exploring Network Security and AI Threats with Crystal Redmann

Friday, 23rd August 2024
Good episode? Give it some love!
Rate Episode

Episode Transcript

Transcripts are displayed as originally observed. Some content, including advertisements may have changed.

Use Ctrl + F to search

0:00

Good morning. You're listening to Breakfast Bites, and I am here today with

0:04

Crystal Redman, who is the operations director from Redman Farms.

0:10

She's a rather inquisitive sort of person and came up with a number of technology-related

0:17

questions that she thought would be helpful to a variety of business leaders.

0:24

Leaders, and since Crystal's in charge of the technology leadership as the technical

0:32

point of contact sort of person for the Redmond Farms,

0:36

she came up with these lovely questions and sent them to me.

0:40

And I'm like, wow, those are really good ones for the show.

0:43

So do you have a question that you want to discuss with me today,

0:49

Crystal, and get to the bottom of some answers?

0:51

Oh so many Felicia okay so

0:55

many okay so randomly just so everybody knows like I'm not like you have not

1:01

told me what the question is so this is going to be totally off the cuff let's

1:05

go that's how our relationship has been so far my one of my biggest burning

1:11

questions is understanding network

1:14

segmentation i kind of

1:18

get it i understand why different there's different wi-fi networks for our business

1:22

all that good stuff but i need to know more like how does it improve security

1:28

are there some common mistakes that people make when they're implementing this what does the.

1:35

Typical person need to know? That's really, it's an excellent question.

1:40

And the reason I think it's such an excellent question is because network layer

1:44

security is the foundation.

1:47

It's absolutely the foundation. It's like when you're doing a house,

1:50

you have to have a lovely hole.

1:53

The hole has to be compacted and you have to have a lovely concrete foundation in there.

1:59

And it has to be all level and square and everything.

2:03

And if that isn't right, then everything else doesn't work quite right.

2:08

So the network is what everything else runs on.

2:11

You need that network to be correct. And network segmentation,

2:15

specifically micro segmentation, is actually better.

2:19

There's a wide variety of ways to do it. And I like to focus on things that

2:26

are the lowest total cost of ownership in terms of methodologies and things

2:30

that are the most sustainable.

2:33

So that means that that level of security or that type or that approach can

2:39

be attainable for even a very, very, very small organization of,

2:45

let's say, a business broker who works out of their house, a single individual.

2:51

So that means we can bring enterprise-grade security-level approaches to even a one-person office.

2:57

And that kind of approach is very important because a lot of the other network

3:02

segmentation approaches are economically infeasible at that scale.

3:09

Okay, so the way to think about segmentation is you are going to treat and classify

3:16

your jewelry box different than your underwear drawer,

3:23

different than your bookshelf, different than your refrigerator.

3:28

These are all different types of assets that you're storing in these different places.

3:34

And so you need to treat them differently. I mean, nothing's going to end well

3:38

if you take your underwear and your jewelry and your books and your food and

3:43

you stick them all together. These are just, right? I mean, these are different things that have different needs.

3:49

You know, the books need to be on the bookshelf with their other books and they

3:54

can't be having milk next to it, getting it all wet.

3:58

Same with your underwear, you know? And jewelry, we typically want that to be

4:03

a little more protected. And so the idea of micro-segmentation at its core is a concept that says we're

4:14

going to understand the assets, and then we're going to enable the assets to do what the assets need to do.

4:21

Like a TV on a guest network, for example, needs to be treated differently than

4:27

a corporate laptop. top. But that's different than a printer.

4:31

And that's different from a switch.

4:34

These different things have different requirements.

4:37

And so if you create a security zone profile sort of concept around these and you segment them,

4:45

it becomes very easy to create a security profile for the segment that the printers are on.

4:52

You can say, well, my rule is I'm only going to put printers Now I can create

4:58

a security zone profile that protects those printers from being tampered with or hacked externally,

5:08

but also helps those printers avoid data leakage.

5:13

Like a printer, for example, could have, let's say it got a piece of malware

5:17

on it, which is totally possible. That can happen. Right.

5:21

But then everybody that's printing, every one of their print jobs might get

5:25

leaked to some unauthorized parties, and that's highly undesirable.

5:31

And so as a result of trying to avoid a situation like that,

5:36

you need to restrict what the printers can communicate with and thus restrict

5:43

what they can send data to.

5:46

And now some devices like a TV, that's what you would call an IoT device.

5:52

The IoT devices, you try your best to create a profile of what they can communicate with.

6:00

But some devices are just really poorly engineered.

6:04

And most TVs, I would say, are the antithesis of anything that's able to be secured.

6:11

But one of the ways that you keep corporate devices from getting ransomware

6:18

is to restrict what they can communicate with.

6:21

Well, if I had to go to the level of effort of mixing TVs and printers and switch

6:32

interfaces and telephones and surveillance cameras,

6:37

you know, and company computers, right? Corporate laptops.

6:41

Oh, and guest devices. If I had to do that all on the same subnet,

6:45

do you think I could protect anything?

6:48

Probably not. Right. It becomes very difficult if it's basically practically impossible.

6:54

It's impossible to do it at an economically viable level because you can't actually

7:02

know what traffic is occurring and what IP address at what time.

7:10

So there you go. I mean, most of IT, or I should say IT security,

7:13

is this balance between trying to protect assets, but still facilitating the

7:19

functionality that someone legitimately is authorized to do.

7:24

Notice I didn't say need, right, because it's not up for employees of an organization

7:28

to discern what they need. It's up for the security management team to listen to them and say,

7:37

okay, I understand you want to do that. Now, what does company policy say about that?

7:42

And then let's see if maybe there's a company policy change that needs to happen.

7:47

I mean, a big example of that would have been when ChatGPT came out.

7:50

I think a lot of organizations failed to appropriately have an AI usage policy

7:58

before they allowed AI to be accessed in that way in their organization.

8:04

And really, they just introduced a whole lot of risk to the organization.

8:07

And network segmentation would have been a very easy way to turn that off if

8:13

it was allowed to be turned off. But unfortunately, most organizations don't function that way.

8:19

And without a governance system like a policy and without leadership who are

8:25

thinking first about risk management,

8:29

then those things tend to go nowhere. So what do you think about that?

8:34

As an answer. Does that help you understand these?

8:37

Absolutely. I'm really glad you said AI because that reminded me of another

8:43

question I have for you. Okay, go for it.

8:48

How do you see AI being used by cyber criminals in the future? Are there.

8:56

New security threats on the horizon that we should be aware of that are coming. How serious is this?

9:02

I think it's wickedly serious. And I mean, I've actually already seen quite a lot of it's come out.

9:10

So one of the things that's super duper cool is we have a system called the

9:16

Breach Prevention Portal System.

9:18

And it's effectively a large scale on demand training platform,

9:24

training and assessment platform for individuals.

9:28

We can deploy it for residential home users and one-user businesses to large

9:34

businesses. It really doesn't matter. It's really cool because we can just provide it to everybody.

9:39

One of the things that comes out of that sort of interaction is that becomes

9:44

a way to provide the answers to those very questions

9:49

to the staff at an organization across the board and to enable personnel managers

9:58

to know that the staff have taken that class or those courses,

10:04

maybe it's a couple courses, and there's always an assessment at the end of it.

10:08

And how did they do on the assessment? Did did they complete the assessment?

10:12

So I've already seen where the criminals are using the AI.

10:18

And I have to mention something else that I think is quite relevant here,

10:23

is that we don't, on our company website,

10:27

we don't list the employees of the company with their photos and their names

10:34

and their fun proclivities or whatever.

10:38

And I have seen organizations do this and it's typically when those organizations

10:43

are driven by marketing people instead of operational security people.

10:48

That type of information is absolutely harvested by the criminals.

10:51

So criminals will also buy all kinds of people finder lists.

10:57

So using something like a service like Abine's Delete Me can be exceptionally

11:02

useful to an individual to help reduce the number of lists that they're on.

11:08

That service is very economical. And again, that's Abine Delete Me.

11:11

I really enjoy that service. I think they do a good job at that company.

11:15

And the criminals are collating all this data using AI.

11:20

They're also drafting messages now that are less prone to have spelling and

11:26

grammar mistakes because now they're not having to do it.

11:29

They're using AI for that. that they're utilizing AI to do deepfakes.

11:35

And that's the one that is really the giant meat and potatoes,

11:40

is the biggest bang for their investment

11:45

is when they're using AI tools to perpetrate deepfakes on people.

11:52

Because let's imagine that an organization didn't have their protocols together.

11:57

And if the person who or multiple people potentially who could be authorized

12:03

to do like a large wire transfer.

12:07

Well, what if they were not immune to getting deep faked and wire frauded?

12:15

Oh, this is where the problem comes in. Did you see the article about the guy

12:21

who was deep faked by like seven?

12:25

He had seven people at his company that were deep faked and he was on the receiving end of all this.

12:31

I think they had a conference call and he was the only, this was a video conference call.

12:37

And he was the only legit dude in that meeting. Did you see that article?

12:41

No, I did not. You'll have to send that to me. That sounds horrifying. I was scared.

12:47

I'm terrified now. It's pretty wild.

12:51

So imagine this dude. This guy was like the ACH wire transfer guy at a bank.

13:00

And he got deep faked in a conference call where there was like seven other

13:06

people in This was a video conference, right?

13:09

How long did he stay on this call?

13:14

I don't know. That information was not in the meeting.

13:19

But I think they got like $50 million.

13:23

Oh my gosh. Yeah, it was pretty crazy.

13:27

Anyway, so this is exactly what AI is doing.

13:31

So, right, if AI is that good now that it can do a video conference call deep

13:38

fake in order to get this guy to make it like a very large wire transfer,

13:44

this is exactly why people need, like, everybody needs this kind of training

13:50

that we have that's available.

13:53

I mean, it's really good training we have. And I've had nothing but very positive.

13:59

Commentary from people about like, yeah, you know, it's really cool. You got AI training now.

14:03

And, and that's really relevant. It's really practical and easy to understand.

14:07

It's like, yes, because it's a big deal. It's a big deal. So what do you think

14:11

about that as an answer to your question? Wonderful. Your answers are always wonderful. So we have AI helping cyber criminals just go wild.

14:23

You don't even need that much knowledge, it seems. You just have AI to help you out here.

14:28

I assume that means malware is probably evolving rapidly?

14:34

Right there yes i

14:38

i think we adapt our strategy to counter like

14:41

these emerging risks well that yeah again

14:44

another fantastic question you have here you know i was thinking about your

14:47

your malware point it's become so super duper easy for bad actors to there's

14:57

kind of like two things that they're doing and i'm sure that i'm i'm underselling

15:00

it right one thing that they're doing is like scenario planning.

15:03

Oh, if I ran my malware and I did this, what's the probability of penetration

15:09

into said company or theft of data or whatever, right?

15:15

So they're now able to do much more sophisticated scenario planning,

15:19

which helps them tune and refine their attack.

15:24

The second huge piece is they don't even have to be able to write code anymore.

15:29

I mean, I mean, this used to be where malware coders used to actually have to write code.

15:34

Now, granted, for probably at least five to six years, there have been ransomware

15:43

rootkit sort of things you could buy on the dark web.

15:47

So this is you got some malicious actors who are coding wizards and they're

15:54

out there and they're basically building like a ransomware kit.

15:57

Kit and you could take a 10 year

16:00

old and some people have done this they'll take a 10 year old put

16:03

them at a computer and say okay use this application

16:06

to go make some ransomware and so here just a 10 year old with a computer and

16:12

this like ransomware wizard generator tool rootkit thing can go and generate

16:19

their own variant and so this is a key piece here is that

16:24

when there's, I believe there was like 100,000 new pieces of malware every hour around the world.

16:33

I mean, there's some sort of just completely outrageous statistics like that.

16:36

And so that's why signature-based...

16:39

Detection isn't that great anymore.

16:43

I'm not saying that people should stop using it. I'm saying that if they're

16:46

using a tool that is centric around that type of functionality,

16:51

it's going to be deficient in its capability set.

16:55

So what we use is we use something that's really a zero trust approach.

17:00

And that's where you really got to get to at the point with malware.

17:08

I could go back at least a decade and

17:11

i can tell you that even a decade ago if a

17:14

p if a computer got some malware on it

17:17

it was done absolutely done i

17:21

mean there was like no removing it so the

17:24

whole thing when people buy they pay the

17:27

the ransomware guys and then they get their you know decryption keys i mean

17:33

i just want to laugh at that type of stuff from the perspective of saying that

17:37

if you think If you think you're going to get data back that actually has integrity

17:42

that you can count on, I think that's a naive thing.

17:45

And if you also think that you're going to get usable systems from paying that

17:50

ransomware, I think that's also a very naive approach.

17:54

Because even 10 years ago, there was really no viable method for correcting

18:00

a computer that had been something malicious happened to it,

18:06

other than you have to wipe the whole thing.

18:09

I even saw, oh my gosh, I can tell you it was the year 2004.

18:13

This was a super long time ago. It was 2004 when I saw partition persistent malware. Malware.

18:23

So if you took a computer and had this nasty thing on it, you rebuild the whole thing.

18:28

And well, shortly thereafter, it was getting something nasty, infecting it again.

18:35

What was the one thing that continued to needed to be done there?

18:39

Well, it wasn't just format, the hard drive. It was all the partitions had to go too.

18:45

So this is something that you have to realize.

18:48

And sometimes you have to question, hmm, has somebody contaminated the BIOS on the motherboard?

18:55

Has it infected a USB attached device like a keyboard or a mouse?

19:03

You know some of these monitors now have like an integrated docking station

19:08

in them that has its own brain chip effectively it has its own motherboard right

19:12

so you're now in the world,

19:17

where you have to assume that everything is a threat and you have to use dynamic

19:22

live updating databases so we're talking about zero trust i have to protect my monitors now too.

19:30

Yes, yes you do.

19:33

I mean, you've always had to protect your monitors from nasty power surges,

19:39

but now monitors have firmware in them. Did you know that?

19:45

No, my mind is right now.

19:49

Okay, so I thought I would just plug it into the wall for a power strip,

19:54

and then you have that HDMI cable, you know, harmless.

19:58

We're good. I don't think about anything.

20:01

No, the HDMI cable is bidirectional in its communications capabilities.

20:08

And so the monitor has firmware, and something has to be able to update the firmware.

20:15

So So there are various communication channels.

20:20

Whereby a computer could do a firmware update, legit or otherwise, to that monitor.

20:28

Actually, this is exactly why I have a very, very strong love for mature and

20:38

enforced procurement policies.

20:40

If you have a procurement policy that does not allow your team to go down to

20:47

random store and procure things like a USB charger, in fact,

20:53

well, I mean, shoot, you've seen these business,

20:55

you've seen these news videos you have to have.

20:59

I mean, I feel like they're like all over the place where is somebody saying,

21:03

you know, do not plug into your mobile phone, you know, a random power cord.

21:10

I didn't sleep for a week after I read the article about these USB phone chargers

21:16

and what's on that could potentially be on them and the threats they pose after.

21:21

Wow. My mom was pwned then.

21:24

Oh, yeah. Yeah, well, I mean, this is one of the reasons why we use wireless

21:28

phone chargers, because that's a surefire way to be successful with that.

21:35

Rather than telling somebody, hey, only use the phone charger that came with

21:39

your phone, I'm saying, hey, don't use the wired phone charger at all.

21:43

Use a wireless phone charger. And they're like, oh my gosh, I got to buy phones that are capable of that.

21:49

And I'm like, the quantity of phones nowadays that can't do that is getting

21:53

even smaller and smaller and smaller. So.

21:57

Bottom line, let's go back to zero trust.

22:01

Zero trust is effectively saying we have to assume that everything is malicious

22:06

until it's been inspected. And this isn't a matter of scanning. It's a matter of do I know you?

22:12

So there's a classification process. Let's say there's an unknown thing that's

22:16

unclassified and it is attempting to do something on a computer.

22:22

Well, the protection tools will grab that thing and they will upload it into

22:30

a virtual machine in the cloud.

22:33

They will execute that thing on the virtual machine in the cloud.

22:37

So this is a terminology we call detonating. You know, it's detonated in a virtual

22:42

machine in the cloud, and then we get to find out what happens.

22:47

Does it hack that virtual machine or not?

22:51

And so this is a way of doing behavior observation and outcome observation.

22:57

Conservation there are also those sorts of processes that

23:00

are going on all the time like a great example is if you have

23:03

really good zero trust threat protection then it's

23:07

going to look at it and say oh we have microsoft

23:11

word is invoking a

23:14

powershell instance what do you think about

23:16

that crystal do you think that word should invoke powershell no

23:21

why would it need to that's exactly

23:24

right so it smells like something malicious

23:27

doesn't it okay it's so

23:31

zero trust is like we're we're gonna do we trust word yes we've classified microsoft

23:38

word but now we're still gonna watch it if microsoft word is doing something

23:44

weird then we have to think that word even though it's a legit tool.

23:51

Maybe it has a malicious plugin embedded in it or something.

23:54

This is no different than if you look at like a web browser,

23:56

take Edge as a web browser.

23:59

It can have a malicious browser plugin. And then that malicious browser plugin

24:03

can do nasty, naughty things.

24:07

So we got like about four minutes left.

24:10

So why don't you give me your next kind of small question if you have one? Small? I don't know.

24:19

Antivirus software. Now, I'm thinking back to the days when I first got a computer.

24:23

You know, you have that notency when you get pushed in and you're,

24:26

as far as everybody knew, everything was great. We're protected.

24:30

I guess, how has that changed since then? And how machine learning,

24:36

has that impacted AI machine learning, the effectiveness of this antivirus software? Is it effective?

24:43

Is it worth it? I don't know. Well, I don't even like to use the term antivirus anymore.

24:49

In fact, I pretty much stopped. As a modus operandi, I'm only utilizing the

24:55

terminology, you know, zero trust threat protection or, you know,

25:00

endpoint protection detection and response. Response and the reason is because

25:06

these things really cannot be unmonitored anymore

25:09

so we do knock sock and mdr and without killing you with acronyms it's just

25:16

basically security monitoring security response actively 24 hours a day now

25:22

we only do that for the clients that subscribe to that service but.

25:27

But that's like, you know, having a little monitor, a protection system,

25:34

like having your own cop that's running around with you all the time.

25:37

Like, you know, the cop's going to watch who's coming in on your zone and should

25:42

you get to go visit Crystal or not. And then how is somebody interacting with Crystal, right?

25:49

That's kind of like having your own personal cop. But we use a lot of machine

25:55

learning and AI for that to keep costs low.

25:59

And the system is really incredible.

26:02

And as a result of that, we have not had a single client get breached who was

26:09

under our full management. The only times we've actually experienced issues is when, you know,

26:16

other people are involved that don't have the same level of training or vigilance on this matter as we do.

26:23

But I don't really use the term antivirus anymore, just simply because antivirus

26:27

represents approximately maybe 10 to 15% of the kind of protection capabilities

26:34

that you need just for your individual computer,

26:37

right? That was my next question.

26:40

Is it enough for personal use? Do I need to supplement with other tools?

26:44

Something like I certainly do. Yes. And so we get to the question of, is this economically viable?

26:51

And I really feel like so many people just think about this wrong.

26:57

They think, oh, well, this is only just for my home use. And it's like,

27:01

yeah, but it's your bank account. You know, it's your identity. How many people do their tax return on their home computer?

27:10

I mean, how many people are interacting with their bank account and their bills

27:14

and their credit card and all of their personal photos and their personal files?

27:20

I mean, these are, you know, the family jewels.

27:23

I would think that they would be looking at this and going like,

27:27

well, I really need some enterprise grade protections here because that's how I think about it.

27:32

I'm like, how much time am I? Go ahead. As a normal everyday person,

27:37

I was guilty of this before. I had no idea that all these threats existed that I was completely naive to

27:47

just go on Amazon, buy a monitor, buy a USB charger.

27:51

Who cares where it came from? Not even looking at the seller.

27:53

Did they just open a shop? No idea. Just purchasing because, well, I have no idea the risks that are out there.

28:01

I don't know the big bad wolves out there. I do now.

28:04

I think a lot of people don't know or they don't care to know.

28:08

Yeah, brands definitely matter. I mean, I would only ever personally purchase

28:13

a SanDisk or a Western Digital. That's it. And there's reasons for that.

28:19

And the brand matters.

28:22

For me, the whole thing around that is that I think that their firmware process

28:26

is done really good and it makes those devices harder to hack. So that's cool.

28:31

Well, we're out of time. Thank you so much for joining me.

28:34

I hope that you will come back and ask me some more of these questions because

28:39

I think your questions are fantastic. Oh, absolutely. I promise I have a lot more for you. Okay. All right.

28:46

Talk to you next time. All right. Thank you. Bye.

Rate

Join Podchaser to...

  • Rate podcasts and episodes
  • Follow podcasts and creators
  • Create podcast and episode lists
  • & much more

Episode Tags

Do you host or manage this podcast?
Claim and edit this page to your liking.
,

Unlock more with Podchaser Pro

  • Audience Insights
  • Contact Information
  • Demographics
  • Charts
  • Sponsor History
  • and More!
Pro Features