Episode Transcript
Transcripts are displayed as originally observed. Some content, including advertisements may have changed.
Use Ctrl + F to search
0:00
Alexander Culafi: The big use case for AI, I think is just
0:04
conceptually minimizing the grunt work that people have to
0:08
do. Antone Gonsalves: Hi, and welcome to Tech News this week.
0:13
I'm your host techtarget, editorial news director and tone
0:18
for Sophos. On today's show, we'll discuss the highlights
0:22
from the HPE Aruba atmosphere conference in the RSA Security
0:26
show. We'll also explain why the Federal Trade Commission has
0:30
taken a hardline position on the use of AI in business versus
0:35
Aruba had its show in Vegas this week, the networking division of
0:39
Hewlett Packard Enterprise launched a redesign of his
0:43
popular network management platform Central, here to tell
0:47
us about the changes and other conference highlights is Bob
0:52
Lella. Bet. And analysts with tech targets Enterprise Strategy
0:56
Group. So what's new in central and how dramatic is the change?
1:02
Unknown: Yeah, it's a fairly significant improvement, great
1:05
user interface, hierarchical approach, kind of a solar system
1:09
approach where you can drill down into different layers. And
1:12
it's much more, I guess, easily organized, all of the end users
1:16
that I've talked to, when I've caught them in the halls or having meals and so forth, have been really impressed with it,
1:21
they're looking forward to using it. But for those who don't want
1:25
to, they've also initiated an easy toggle switch where you can
1:28
go between interfaces. So it's that ability to leverage the new
1:32
technology. But if you're not comfortable with it, you just
1:34
press the button, and you're back to your usual usual
1:37
comfortable interface. But I think like I said, they're,
1:40
they're trying to make these changes to accommodate the more
1:44
highly distributed environments, more complex environments, and
1:47
making it easier to be able to quickly go through and diagnose
1:51
problems to be able to have that full end to end visibility
1:54
across these large environments, as they now are, you know, going
1:57
between the data center all the way to the edge, right with all
2:00
their solutions. Antone Gonsalves: So So what's changed? Well, why introduced
2:05
this, this more graphical visualization interface? I mean,
2:11
why do it now versus I don't know, three years ago, four
2:14
years ago, what's changed in the in the in the market? Yeah,
2:20
well, like Unknown: I said, I think one of the biggest reasons for doing
2:22
this, and especially the hierarchical piece is the fact
2:25
that the environments are getting that much more complex,
2:28
right? Our research has shown that or more and more organizations are moving applications to multiple public
2:32
clouds, more applications are moving to the edge, right,
2:36
you've got all the remote workers, things like that. So
2:38
things are far more distributed and more complex and trying to
2:41
be able to display that complete end to end interface in one
2:44
image on the screen is going to be highly complex. So being able
2:48
to break it down, and then have it be able to alert and bring
2:51
you to the areas of concern more quickly, are all things that are
2:54
going to help drive operational efficiency, which is what is
2:57
sorely needed for these IT organizations to deal with this
3:00
complex environment. Antone Gonsalves: Yeah, I think the majority are a large portion
3:04
of Rubezh. customer base are universities. So I was thinking
3:11
that possibly these changes in Central is a is in recognition
3:16
of the fact that the IT talent University is usually very few
3:20
people. And they're not high level engineers, you know,
3:24
necessarily, Unknown: that's it's, I mean, the hard part, I think overall,
3:28
is just there's just a lack of skilled IT resources in general.
3:32
At the same time, the IT environment is getting more
3:34
complex. And so in addition to the skilled resources, it's also
3:38
even for the skilled resources, is there's a lack of time to be
3:42
able to accomplish all of these things. So So in addition to the
3:45
new user interfaces, you see them also doing a lot of
3:48
enhancements to their AI operations as well, to be able
3:51
to, you know, eliminate having to do those mundane routine
3:55
things or to be able to accelerate troubleshooting. So
3:58
you're not spending hours trying to find the problem, you're just
4:00
able to see it and fix it quickly. So I think it's a
4:03
combination of those things. But I mean, I spoke to customers
4:07
that who were running, you know, big stadiums and things like
4:10
that, they were very eager to adopt it and look at it as well
4:13
as you know, the colleges and universities. So I think across
4:16
the board, what I've seen this week is that organizations, you
4:20
know, they need to be able to certainly that shift left
4:22
mentality and be able to move faster, get closer to DevOps,
4:25
and things like that. And so these new interfaces, the the
4:28
intelligence and automation that's being added to them are
4:30
all steps that are being taken by the by Aruba to enable
4:34
organizations to to accommodate those needs and to shift left
4:38
and to drive greater operational efficiencies.
4:41
Antone Gonsalves: Okay, we know what I didn't know they have an
4:43
SD win, as we know, I mean, they bought silver peak few years
4:47
ago, but I didn't. I didn't see any announcements related to
4:53
cloud networking. Do you expect that that topic to be addressed
4:57
eventually by Yeah,
5:00
Unknown: Mike Thanks. So I mean, they've got they had an SD Wan
5:02
solutions part of the SD branch, they then acquired silver peak,
5:05
as you said, that's brought on substantial Connect, you know,
5:08
substantial SD Wan capabilities and footprint. They've now
5:11
acquired axis as well, right, the security side. So they've
5:14
got an integrated sassy offering. And so that's a lot of
5:18
the right dealing with the remote workers and things like
5:21
that and hybrid work. But they are starting to look at you
5:24
know, and there's been a couple of the their sessions here that
5:27
have talked about multi cloud networking. A lot of them have
5:30
been about the SD Wan connecting to the cloud, but I think that's
5:33
an area of of, I guess, investigation for them. And I
5:37
would expect that over time, we'll probably hear more from
5:40
them about the the multi cloud environment as well.
5:43
Antone Gonsalves: Okay. And lastly, you know, they also
5:46
rolled out something that they call the Agile network as a as a
5:50
service. What is that?
5:53
Unknown: Yeah, absolutely. So it's the just take a step back,
5:56
right? HPE, which they are now becoming more tightly integrated
6:00
with, and that was indicated by the new logo. I don't know if
6:04
you saw any of my tweets, but they have. It used to be bigger
6:08
Aruba, small HPE, you know, Hewlett Packard Enterprise
6:10
company, now it's big HPE, Aruba networking, so there's there's
6:15
some tighter integration, but they still retain a lot of their
6:18
their autonomy. But a big piece of that is integrating with
6:21
Greenlake and being able to deliver things as a service.
6:25
Now, what Aruba has seen over the last several years is that,
6:28
you know, one one size does not fit all, when it comes to
6:32
networking solutions, how people want to do it, right. Is it
6:34
going to be fully managed, co managed? Do organizations want
6:38
to have control? Is it subscription? Is it consumption
6:41
based? Right? I mean, there's there's lots of areas, I think
6:43
they have six key components that they use to define what an
6:47
agile NASS environment is. But for them, it's really all about
6:51
giving their their customers choice. It's about how do you
6:54
want to consume this in which format, how much management
6:58
levels, right, and kind of sliding bars across all six of
7:01
these terms, to let the customers define the best fit
7:04
for their unique environment? Antone Gonsalves: Sure. So it's pretty much however you want to
7:10
consume their their products, they'll sell it to you
7:13
Unknown: directly. And so they're not they don't want to force them into this is our fixed definition. And you have
7:17
to consume it this way or the highway. It's much more of let's
7:21
figure out what what works best for you. You know, and
7:23
especially when it comes to pricing, we always say you know,
7:26
surprises are for birthdays, not your your IT bill. So you know,
7:29
do you want it subscription based? Do you want it as a lease? Do you want it as a true consumption, etc.
7:37
Antone Gonsalves: Okay, next is RSA AI. And its impact on
7:41
security was a major theme at the show in San Francisco here
7:45
to tell us why is tick target editorial security reporter Alex
7:50
coffee. So before we get started on AI, what was the mood like in
7:55
this year's show and what stood out to you?
7:58
Alexander Culafi: So this is my third RSA, the first one I went
8:03
to was early 2020. I had been on the security beat, and actually
8:08
in like, traditional tech journalism for a few months at
8:11
the time, and it was also as COVID was happening, because it
8:14
was I think, late February, I think I met you in person for
8:17
the first time at that RSA. And then 2021, there was no RSA
8:22
because of COVID restrictions, folks still figuring it out. I
8:25
still being in the thick of things. 2022 They brought it
8:29
back, but there was like a convoluted but necessary, like
8:33
process for confirming that people were vaccinated. And you
8:37
could feel the presence of COVID over that show. And in part
8:42
because there weren't that many people 2023 has been a big tone
8:47
shift and that it was packed, maybe as packed as 2020 was, I
8:52
don't know the numbers exactly. But it was very fact. Like also
8:56
not very many masks like I had mine on the whole time. But I
9:00
gotta say you were hearing that conference cough after a couple
9:03
of days. But on the security side, it's cool, because people
9:10
are getting out there again, holding meetings, you see life
9:13
coming back to normal, and you see people, you know, excited to
9:17
pitch stuff again, like it's not low energy, like last year was
9:21
and I think that part was really, really nice.
9:24
Antone Gonsalves: Sure. And what stood out for you on the tech
9:28
side, Alexander Culafi: who is AI? Like I know that after Microsoft
9:32
sort of announced their multi billion multi year investment
9:35
into chat GPT publishing or open AI a few months ago that like aI
9:40
was going to be the big thing across all the enterprise
9:43
departments that we work on a techtarget editorial, but
9:48
Security I was surprised to see the the prominence of the AI
9:56
that was being sold on the show floor some of it a little
9:59
buzzword already may be a little misleading if I looked into it
10:02
further for some of the smaller vendors, but you also had Google
10:08
and IBM announcing new AI powered suite versions of their
10:14
of some other security products along with some other vendors
10:17
announcing some security products, but like aI was fully
10:21
in the air, Antone Gonsalves: right, so So how is how would AI in general,
10:27
you know, change cyber and cyber security? What are the
10:31
components? And Alexander Culafi: so I talked 567 people this week at RSA, and
10:36
I asked them that very same question, like, what is this
10:40
moment of AI and cybersecurity? Is it new technology? Is it a
10:45
big buzzword? Or is it both? And they're like, yes, it's both.
10:49
And so part of it is messaging, like chat GPT is gone into
10:54
public's hands. And now people are way more receptive to AI as
10:59
a as a concept. The public is and as such, what I've heard is
11:05
that enterprises are feeling more empowered to start selling
11:08
AI, or to like put it let's say, front and center, even though
11:11
like vendors have been using AI and machine learning for some
11:15
cases for decades. So it's not exactly new technology. And a
11:19
lot of the technology that's coming out isn't strictly new
11:22
technology, but refined technology. On the other hand,
11:26
there are some cool applications that are starting to emerge more
11:30
like I think Google and IBM, they can prioritize threat
11:33
alerts. So it can sort of, if there's multiple points of
11:39
concern in your network, it can prioritize and act on them in
11:43
order, there's automated threat hunting. So it identifies the
11:47
software identifies the threats, and then sort of appropriately
11:52
goes after it without you needing to do much, which is
11:55
kind of pre existing but refining technology. However,
11:58
the big use case for AI, I think, is just conceptually
12:04
minimizing the grunt work that people have to do. So like I so
12:10
for instance, in our job as journalists, like, I'm, I'm even
12:16
28, I'm old enough to remember like recording things, and then
12:19
transcribing by hand, a 40 minute interview with something
12:22
like otter AI, you plug in the mp3, you can pull quotes, listen
12:26
back, get exactly what you need. And like an eighth of the time,
12:30
and security has its own versions of that where the data
12:33
processing is getting faster. The analytic stuff is getting
12:37
faster, and security professionals can do the
12:41
important stuff and focus on the more important stuff quicker.
12:45
And I think that sort of where the exciting stuff is, is that
12:49
all the annoying work, or the stuff that's like that used to
12:54
be more granular is starting to get automated and like code, for
12:59
example, it can like, I've heard that like chat GPT, or like open
13:04
AI, stuff can can generate like pretty good code, or check it
13:10
pretty well as well. And I think that automation, and that like
13:14
speeding up, what was once completely manual process. I
13:18
think that's the gonna be the exciting stuff right now across
13:21
the board, including security, if that makes sense.
13:25
Antone Gonsalves: Yeah, it does. It seems like the initial use of
13:28
AI is all about efficiency and saving thing and people's time.
13:35
I mean, it'll be interesting to see, as the technology matures,
13:40
how it gets more, you know, more Incorporated, it goes beyond
13:44
just just efficiency, you know, like, helping people like
13:49
actually making decisions in terms of how to how to protect a
13:54
company, you know, launching some type of defensive software,
13:59
you know, in the case of an attack, you know, without human
14:02
intervention, that type of thing. But, but it's it but it's
14:07
nevertheless it's, it's, I agree, it does seem to be it's
14:11
real. And it's just a question to see and how it evolves, you
14:14
know, as the years as the years go by. Finally, the Federal
14:20
Trade Commission is ready to pounce on companies that misuse
14:25
AI. Here to tell us more about the FTC warning is tech target
14:31
editorial government reporter MacKenzie Holland. So what did
14:35
FTC chair Lena Kahn mean when she said there is no AI
14:40
exemption for business? Unknown: consumer protection laws, like the Federal Trade
14:45
Commission Act itself, were signed into law in 1914.
14:50
Obviously, artificial intelligence didn't exist then.
14:53
And there's been a lot of concern that those laws that
14:56
were signed into law so long ago don't cover now New technologies
15:00
like artificial intelligence. This week, the Federal Trade
15:03
Commission, along with the US Department of Justice, or Rights
15:07
Division and other consumer protection agencies, essentially
15:11
came together to take the stance that these laws do apply, no
15:16
matter the technology, these consumer protection laws apply.
15:21
So the laws aimed to stop bias and discrimination. It doesn't
15:24
matter if it comes from a real person behind the desk or an
15:27
algorithm on a computer. Essentially, these laws apply,
15:30
no matter how old they are, there are ways to make these
15:33
laws apply. And that's what FTC Chairman ICANN meant when she
15:36
said that there's no AI exemption. Antone Gonsalves: And what are the potential problems that FTC
15:41
is worried about? You mentioned discrimination. There. Other
15:46
others? Unknown: Yeah, the problems are that biased algorithm, because
15:51
there are studies that reveal that AI algorithms can be biased
15:56
in depending on the data that they're fed, and other factors
16:00
included in there, and that they could negatively affect large
16:03
swaths of people, but especially minorities and those with
16:07
disabilities, that could be denied housing, credit, even
16:11
employment. For example, the company uses a hiring algorithm
16:16
that the employer programs to select candidates similar to the
16:20
existing employment base, the algorithm could reproduce the
16:25
traits of that group. So if it's an employment base of 70%, white
16:30
males, the algorithm could look for similar traits, and preclude
16:34
something like accommodations for disabled persons. So there
16:37
are plenty of examples out there where these algorithms could
16:42
negatively impact people looking for jobs, people looking for
16:47
lending opportunities, looking for housing, depending on the
16:51
traits that these algorithms do take into account. So there are
16:54
plenty of issues on hand with this.
16:57
Antone Gonsalves: I think trends when it comes to AI,
17:02
transparency is still something that I think is a problem is we
17:07
don't really have transparency in these algorithms. How they
17:11
make, how the how they make the decisions that they make, you
17:14
know, I did talk to an attorney with a very large law firm, last
17:21
week, fully laden. And he made the point that when it comes to
17:25
AI, even though you're buying the software from somebody,
17:31
you're still responsible for, for the output, not the vendor,
17:37
the vendor who sells it to you, which I think is very
17:41
interesting. You're pretty much no matter what happens, what
17:46
abuses result from the problems with the software, as the as the
17:52
user as the company, you're still responsible for it. So
17:56
alright, so it sounds like the legal teams and companies are
18:01
going to be are going to be Yeah, yeah. And
18:04
Unknown: you know, because because these enforcement agencies do typically go after companies using these products
18:09
to hold them accountable. But during the press briefing this
18:12
week, where they kind of made this stance on holding AI
18:16
accountable. FTC chair Lena Khan did point that out, she did
18:20
recognize that these tools are often released into the wild
18:25
lawmakers have to do cleanup after the fact. And so that's
18:29
what she did say that the FTC is going to be taking a second look
18:32
at where they enforce the law. She said the FTC has put out a
18:36
notice to the market clarifying that if companies are building
18:39
and producing AI tools that are designed to deceive, then those
18:43
companies could potentially be on the hook and subject to
18:46
liability for those products. So it's not something that the FTC
18:49
goes after now. But it's definitely on their radar, they
18:53
recognize that it's an issue with these products coming from
18:57
the the company that makes them going out into the wild where
19:00
other companies use them, that that could be the source of the
19:03
issue. And so I think they do recognize that that's a problem.
19:07
But they're just they don't have the tools to go after it yet. So
19:10
they're, they're recognizing that working on that, surely, I
19:13
Antone Gonsalves: would expect companies and probably doing it
19:17
now that when they do buy software, that's heavily
19:22
dependent on AI, that they will ask the vendor, they will
19:27
require that the vendor, tell them how the algorithms work, be
19:32
transparent, at least in contract negotiations and
19:37
Unknown: have risk management policies in place because there
19:39
are those risks with AI that they have to mitigate and try to
19:42
work around. Antone Gonsalves: Absolutely, absolutely. And then there's
19:46
also the question of you know, you have risk management and
19:53
they would want some kind of guarantees that the that the
19:58
output or the data as I was thinking about the data that
20:01
goes into the software I think companies will want to possibly
20:05
use not just data from the internet, but cleaner data, you
20:10
know, information that they have that they can control so that
20:14
they can have a better chance of getting the output that they
20:17
want. You know, exactly. Yeah. So there's a lot a lot happening
20:21
here and I'm sure you will be watching closely. So. Alright,
20:25
so that wraps up this week's show. Thanks for watching and
20:29
enjoy the weekend. I'll see you next week.
Podchaser is the ultimate destination for podcast data, search, and discovery. Learn More