June 16, 2023 • 23 mins
Each week, TechTarget Editorial's news team and industry experts provide candid insight and analysis of the biggest IT news headlines of the week. Today's show covers the generative AI plans within Cisco, Oracle, and Salesforce and the European Union's big step toward AI regulation.
Watch Tech News this Week on the TechTarget News Bytes Channel:
Stories featured on Tech News this Week:
Oracle plans to provide generative AI services with Cohere
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Makenzie Holland (00:00):
I think the most significant part of that
(00:01):
legislation is the way that aclassifies AI into risk.
Antone Gonsalves (00:06):
Hi, welcome to Tech News this week. I'm your
host TIG targeted to real newsdirector Anton good Sophos.
Today we'll discuss how Ciscoplans to become a stronger
security provider, the state ofAI regulations in Europe and the
US, and the latest generative AIintroduced by Oracle and
Salesforce. at Cisco Live lastweek, security was as much a
(00:30):
topic of conversation asnetworking, here to tell us
about the networking giant'slatest security products. And
the strategy behind them isMelinda Marques and analysts at
Tech targets Enterprise StrategyGroup. Welcome, Melinda.
Melinda Marks (00:46):
Hi, Anton. Great to be here.
Antone Gonsalves (00:48):
Great. Great to have you. So Cisco, how is
Cisco's advancing theintegration between security and
networking in its latest productannouncements from from Cisco
Live last week?
Melinda Marks (01:04):
Yeah, I think they're making a really smart
point about how importantsecurity is to conductivity and
networking. So they made it amajor theme at the conference
last week, you know, Cisco isn'treally known as a security
vendor, or that's not the mainthing that they're known for.
But they made a point of talkingabout how that's the number one
(01:24):
topic for customers. So anytimethere's a connection, it needs
to be secure. Their customersneed to be able to use Cisco
products easily, while havingconfidence that it's secure. So
there are a lot of greatsecurity initiatives that they
announced last week, theyannounced secure access, which
is to make sure that anybodytrying to connect and again,
(01:47):
this is a big point for modernpressures of digital
transformation, you have a lotof remote work, hybrid work. And
they announced that theirsecurity service edge, their
sassy solution really helpssimplify access. So if employees
are connecting from any locationfrom any device, that could do
(02:07):
so and work productively. Andthen on the back end, there's a
easier way to manage thataccess, see where people are
connecting. They also announceda really cool collaboration with
Apple to incorporate zero trustnetwork access for a seamless
native experience for Mac users.So those are really nice, nice
(02:29):
things to see that most peoplecan understand the need to
access and be able to accesstheir work applications
securely. So I think thiscommunications pattern that
they're doing have, make Ciscomore associated with security
was really strong at that event.They also had some great
(02:50):
speakers they had the NFL wasvery present at the show, they
had their party and Allegiantstadium, so they were talking
about how you can have thesegreat experiences securely. So
if you're at a ballpark or astadium, you can go order your
food, you can do everything withyour mobile app, you don't have
to worry about carrying yourwallet. So these are all things
(03:13):
that Cisco's really emphasizingas ways that they can make
modern life better, moreconnected, while making sure
that they're fully thinking ofall the security implications,
so that you don't have to, andyou can just go about what you
need to do.
Antone Gonsalves (03:29):
Right. And they also introduced generative
AI within within their securityportfolio, right? I mean, how's
it going to be used in thecompany's new policy assistant?
Melinda Marks (03:38):
Yeah, and I love that because this is a big
challenge is there's a lot oftedious manual work that's
involved. Cisco made a big pointabout responsible and ethical
usage of AI. There's a lot oftalk right now, with I think,
this year early this year withthe emergence of chat GPT, what
(03:58):
are AI implications, and peoplewant to use it in a smart way to
make things easier and to takeaway manual work and things like
setting policies and settingguardrails. That's something
that takes a lot of manualtedious work. So they're using
their generative AI policyassistant for Cisco secure cloud
(04:21):
so that security and IT adminscan set and maintain security
policies. And this is alsoagain, it's to help them reduce
their manual work. It's to helpmake sure that there's no
duplicated work. These are allthings that really help security
folks, make sure that they canset up policies and easier
scalable way.
Antone Gonsalves (04:42):
Right, and how about generative AI in their
security operations center?That's also an area that they're
going to include the technology.
Melinda Marks (04:51):
Yeah, so they announced leveraging generative
AI for faster threatintelligence, threat detection
and response. So when I'mThere's some issues security
issue, it's really difficult todetect where it is and respond
quickly. So at the end of thisyear and early next year, they
have some new features wherethey have security operations
(05:15):
center assistance, so that youcan contextualize the events and
respond faster, so that you can,you know, get to the problem
faster and, and remediate it.And with those types of tools,
they can better understand whatthe issues are, understand what
their impact is, and make surethat they put in the best
remediation efforts to containthat or stop it as fast as they
(05:39):
can.
Antone Gonsalves (05:40):
Okay, Cisco also launched panopticon, which
brings us into cloud security,right? We're talking about cloud
native applications. I mean,that's, that's new for Cisco.
How will it you know, expand itsits portfolio?
Melinda Marks (05:57):
Yeah. So this was exciting. And I actually got to
speak in their session to talkabout it. So when we think about
the needs for digitaltransformation and moving
workloads to the cloud, there'sorganizations are moving to
cloud native development,because it makes development a
lot faster and more efficientfor developers. And Cisco has
been up to cup and Africa istheir solution for
(06:21):
observability. And visibility,as the name implies, with
panopticon that you can seeeverything. And when you think
about managing security and riskwith the move of workloads to
the cloud, and secure andmanaging security, in general,
in general is always aboutvisibility, and making sure that
you can see into everyconnection, every resource, and
(06:41):
with the dynamic ephemeralnature of the cloud and, and
development resources wherethey're using these. They're
using applications based onmicroservices. So very
distributed, you need thatvisibility to be able to see and
monitor those applications inorder to manage them. And again,
to find security issues and makesure that you can remediate
(07:04):
them. So they have, theyannounced some updates to their
full stack observabilityplatform. And this is integrated
with panopticon. So that's whereyou get that full visibility.
They also have those featuresand capabilities that they had
from AppDynamics and 1000. Eyes.So and they're also using a lot
(07:25):
of their innovation and emergingtechnologies that they found
this as part of their out shiftproject, which is their
incubation engine for emergingtechnologies, including the
cloud, native stack, edgecomputing, AI and quantum
computing. So when you bring allthis together, and you help
people with security, they'rebringing together a lot of
(07:47):
security capabilities that youlook for in a cloud native
application protection platform,which kind of brings everything
together for efficient riskmitigation and application
protection. So now you havethings like, you can monitor the
applications. And you can theyalso have some infrastructures
(08:09):
code scanning capabilities,workload monitoring
capabilities, and applicationand API security capabilities
that they demoed in the sessionthat I was in that show how this
gives people a platform thatties in what what assets they
have in the cloud, how tomonitor the workloads, and then
identify any issues and makesure that they have fast,
(08:31):
effective feedback loops fortheir developers. So when an
issue is found, it canimmediately go back to the
developer, so the developer canimmediate it with speed and
efficiency. So we
Antone Gonsalves (08:42):
look at Cisco and Cisco's portfolio is huge.
We know that we could spend twoand a half hours just going
through all the all the featuresthat it provides. But you
mentioned early on about Cisconot being seen as a security
vendor security provider, atleast not a leading one.
Melinda Marks (09:03):
Networking, yeah, use extensive
Antone Gonsalves (09:06):
portfolio, you would think the opposite? Is it
because Cisco has seen itssecurity portfolio was seen as
something closely tied to itsnetworking and therefore
something most beneficial toCisco shots, enterprises that
are heavily invested in Cisco.And
Melinda Marks (09:24):
yeah, absolutely. And I think this is why I'm
excited that organizations arelike Cisco, where security is so
important that they meet thishead on right. I've been at
technol. I've been on the otherside at technology vendors,
where security is an issue. It'syou know, when you think of why
would people not buy my productsecurity, if it can be a
(09:47):
blocker, they think it'sinsecure. You know, pricing and
other things might be a reasonwhy people might be wondering
what whether they should buy itor if they should choose another
product and I I really like thisboldness of Cisco to just take
it head on and say, No, we'regoing to we're going to meet
these needs and where you wantto be a major security vendor,
(10:10):
we want Cisco Live to be a topsecurity show that was kind of
surprising to me and all theirpress releases. This is a
networking and securityconference, where they're trying
to position it that way. So, youknow, to me as a security person
that thinks that's the mostimportant thing that people
should care about. And it shouldbe an enabler and not a blocker.
It's just really refreshing tosee them put that emphasis. And
(10:31):
then also, because they havethat reach of the types of
products that they have, likethey're known for their network
security, they're sassy, theydid announced some firewall
updates as well. You know, thisreally expands into like the
area that I cover, which iscloud and application security,
and helping with digitaltransformation. And they do have
(10:52):
a lot of advantages therebecause they have the different
components that they need thatyou need to address that well.
So things like securing at theedge for the developers as
they're using their laptops towrite code spanning to the
network spanning all these areasthat you really need to do and
get control of and disabilityhave an order to mitigate risk.
Antone Gonsalves (11:16):
While US lawmakers did or over AI
regulations the European Unionis made is moving toward the
passage of an AI act. Here totell us the impact of the EU's
aggressive action on AIregulation is techtarget.
Tutorial government reporterMacKenzie Hollen. Welcome,
Mackenzie.
Unknown (11:36):
Thank you, Anton happy to be here.
Antone Gonsalves (11:37):
Right. So what's in the AI act? And where
is it on the road to approval?
Makenzie Holland (11:42):
The AI act is legislation that essentially
classifies risk of artificialintelligence and high low and
prohibited artificialintelligence use. So
essentially, lawmakers havedeemed that, you know,
prohibited uses of artificialintelligence include facial
recognition, or using AI toclassify people based on
(12:05):
socioeconomic status or otherpersonal characteristics. So I
think the most significant partof that legislation is the way
that a classifies AI into risk.So essentially, you have
different obligations to complywith based on the level of risk
that your AI is classified at.And so where they're at in this
process is the EuropeanParliament just approved this
(12:27):
version, this draft version ofthe AI Act. Now it goes to the
European Council, and they'regoing to be speaking with other
member countries to finalizethis law and essentially propose
what the final version is goingto look like. And it's already
moving a lot faster thisprocess, then people thought
that it was going to so this isalready expected to be a final
(12:47):
loss expected to be proposed bythe end of the year.
Antone Gonsalves (12:51):
Okay, and what's the what's the impact
potential impact on UScompanies? I mean, what are your
sources telling you?
Makenzie Holland (12:59):
Sure. So it's, there's going to be obviously a
time period where companies aregoing to have you probably two
to three years to comply withthis legislation. So there won't
be immediate impacts. But it'sgoing to be something that, you
know, companies will have toadjust to, because they even
have a section in there forgenerative AI. And that's
(13:20):
something that we've been seeingtaken off recently. So
essentially, companies are goingto have to be transparent about
their generative AI use andgenerative AI is going to have
to essentially tell users ifthis content was aI generated,
and there are other transparencymeasures behind that. So
companies are going to have alittle bit of time, you know, to
comply with this legislation,obviously, once it is approved
(13:43):
and finalized. So it's hard tosay what the immediate impacts
would be, but it's there. Peopleare gonna have some time.
Antone Gonsalves (13:50):
Yeah. When you talk about transparency and
their generative AI us, doesthat mean, they got to hand over
their algorithms to the EuropeanUnion or something, something,
Makenzie Holland (14:00):
you know, it's, again, these this is not a
final form of this law. And it'sa kind of broad section. So I
think it's up for interpretationa little bit. But yeah, right
now, it's just essentially kindof asking for, to tell whether
or not this content is AIgenerated. Have the training
data on hands, essentially, kindof the I don't think they're
(14:23):
asking to turn over thealgorithm itself, but just for
transparency into outputs. Andwe'll see where it goes from
there.
Antone Gonsalves (14:33):
And in the US, so either what are some of the
proposals being brought forthand as to how we would regulate
AI if we ever decide to
Makenzie Holland (14:45):
regulate? Yeah, there are a lot of
different thoughts on that.Following open AI CEO Sam
Altman's testimony back in May,Senator Michael Bennet proposed
a bill to create a federalagency as the typically to
regulate artificialintelligence. That's one thing
that's been proposed. Altmanhimself proposed, you know,
(15:07):
licensing requirements for largeAI models. And then even
yesterday, Senator Josh Hawleyintroduced a bill that would
carve out generative AI fromsection 230. immunity, which
section 230 protects platformsfrom content generated by third
parties. And so they'retargeting AI that way.
Antone Gonsalves (15:27):
And there's also me there are people who
favor the tech industry,regulating itself set their own
standards. I mean, that sounds,that doesn't sound like a great
solution to me, because you are,you're putting the fox in charge
of the henhouse and AI isimpacts so much of our lives. So
(15:50):
it has the potential ofimpacting so much of our lives.
So what are what are the prosand cons of that particular
approach?
Makenzie Holland (15:57):
Sure. So I guess I could touch on the cons
of that first, becauseobviously, we've seen what's
happened with companies and selfregulation, when it comes to
social media platforms.Obviously, there wasn't a lot of
self regulation going on when itcame to data collection,
privacy. And so we've kind ofbeen seeing some issues with
(16:17):
social media as a result of kindof letting companies and do
their own thing with that. Someof the pros, these industry led
standards have underlyingtechnologies like Bluetooth like
Wi Fi, and, you know, there'skind of a universal agreement on
how these kind of tools shouldbe used. So it's, you know,
(16:37):
there are pros and cons to theapproach. I think it's just
going to depend I don't thinkwith artificial intelligence is
going to be left solely up tothe companies there's too much
and what we're seeing here,Europe already with the AI act,
they're not going to just letcompanies do this on their own.
Antone Gonsalves (16:54):
Right. So it seems like the tech industry is
going to have to, and theyalready are worked very closely
with government. And, andhopefully consumer groups to be
involved to to protect, protectthe rest of us. This week,
Oracle and Salesforce made bigmoves in generative AI Oracle
(17:17):
said it would offer its cloudcustomers, startup cohere for
building generative AIapplications, Salesforce has a
different approach. Thecompany's new API's Cloud Suite
includes an Einstein GPT trustlayer that prevents generative
AI models from holding yourcompany's customer data. Here to
(17:38):
discuss the latest moves fromOracle and Salesforce is tech
targeted tutorials AI reportingthe rest of his job. So what
does cohere offer companies thatwant to run AI applications on
Oracle Cloud,
Esther Ajao (17:52):
so are going for air are kind of working together
so that customers can traintheir own customized large
language model. So what articleas well as a melee other cloud
providers but melee are goodselling point. And also
Salesforce says we're going todiscuss later on is this idea of
protecting customers data. Andso what they're offering
(18:16):
together on OCI is that theywill plan on offering like a
service that will not mixcustomers data with Arcos data
or with chorus data. So coreselling point is they are very,
like, your data stays your data.And we are not going to mix that
together, if that makes sense.
Antone Gonsalves (18:35):
So the I take it customers would use their own
data to train the models isYeah,
Esther Ajao (18:43):
yeah, customers will use their own data to train
the model. But with like others,right, you especially the
problem, not the problem. Butmany of the challenges or some
of the drawbacks that customersenterprises have with tools like
open AI chat. GPT is this mixingof data, so to speak, their
concern that perhaps likesensitive data, especially if
(19:05):
you're working in like, thosesensitive fields, like the
healthcare industry, will kindof lead in and you will be their
data will be used to train thesystem. And so cohere,
specializes in not doing that.And so this partnership is kind
of good. It's very good forOracle. So it's good for us. So
Correct.
Antone Gonsalves (19:23):
Okay, what tell us a little bit, a little
bit about the status quo hereand its relationship with with
Oracle.
Esther Ajao (19:30):
Yeah, so Korea is kind of like open AI. But they
work primarily with enterprises.And so they let enterprises use
their own data to train the AImodels. So they're not sharing
the data, the data stays, itstays within the customer stays
within the enterprise. And solast week, they announced that
they have raised 270 million,and part of the fundraisers are
(19:54):
part of the people involved inthat fundraising is Claire, I'm
sorry, it's Oracle as well asNvidia And so Oracle, obviously,
is investing a lot of money intoCO heir as other companies. But
this is a good relationship forOracle, who can't like
necessarily go to my open AI,which already has a partnership
(20:14):
with Microsoft, as you very wellknow.
Antone Gonsalves (20:16):
Yes, I believe Salesforce ventures A, which is
competitors also invested in COhere.
Esther Ajao (20:24):
Yeah, yeah, hearing voices us also invested. So it's
not like coin is like married toProko. They have the investor,
as well as they also have a lotof they also like, I think
they'll also have products thatcool air is like, that
integrates chorus technology.But this is kind of like a
(20:45):
little bit of helping Oracle getinto the large language model,
generative AI market. Obviously,Oracle has a big database, that
will be helpful to Korea aswell.
Antone Gonsalves (20:57):
No, I mean, it would seem that the the approach
that cohere takes, who iscertainly in line with what,
what enterprises want, I mean,I've talked to people who say
that they won't even usegenerative AI, unless the vendor
can show can prove that the datais, is protected.
Esther Ajao (21:21):
Yeah, that's, that's, that's, that is quite
shallow, but that I thinkthere's a lot of things that
enterprises are like, concernedabout is that data aspects,
obviously, like I mentioned, ifyou are in the sensitive field,
you don't want your data to beused to train a model like I
don't want, I don't want my, youknow, like, I don't know, like
my DNA sample, or like myhealthcare information on chat
(21:44):
TBTF, so to speak, like that.And so that was I think, even
just to re emphasize was, thatwas one of the selling point of
their whole Einstein, GBT trustlayer is that you will have that
trusted asset, that trust thelayer that you really need, in
order to so that you can you canknow that your data is secure,
you can know that your customerdata is secure. And it's not
(22:07):
being used in this AI system.And also it's kind of
explainable, you know, wherethat data is coming from, you
know, it's not going into some,some black box or you don't know
anything about
Antone Gonsalves (22:17):
That's right. But Salesforce, Salesforce
Salesforce is approach, I usedto be a little bit different,
because you're not actuallycreating the AI application.
They're gonna give it to you.But they have that control layer
that the data from from fromheading to someone else's
(22:37):
database.
Esther Ajao (22:40):
Oracle and curry I taken their approach of you will
be able to make their own LMmodels, you will have new app
access and controls everythingout to everything. Obviously
Oracle also plans on embeddinglike core technology and models
into their businessapplications. But it also gives
the control to the enterprise aswell as Salesforce is like we
(23:01):
will have that extra layer thatyou need to know that first show
this is really secure. Sodefinitely two different
approaches.
Antone Gonsalves (23:09):
Okay, and when is is coherer available now on
the Oracle platform and whatabout that Einstein GPT trust
layer? Is that available now orin the future?
Esther Ajao (23:22):
I believe it is not available. I believe it is in
the future. I'm not I don'tremember that on
Antone Gonsalves (23:30):
All right, so that wraps up this week's show.
Thanks for watching. Enjoy theweekend.
I think the most significant part of that
(00:01):
legislation is the way that aclassifies AI into risk.
Antone Gonsalves (00:06):
Hi, welcome to Tech News this week. I'm your
host TIG targeted to real newsdirector Anton good Sophos.
Today we'll discuss how Ciscoplans to become a stronger
security provider, the state ofAI regulations in Europe and the
US, and the latest generative AIintroduced by Oracle and
Salesforce. at Cisco Live lastweek, security was as much a
(00:30):
topic of conversation asnetworking, here to tell us
about the networking giant'slatest security products. And
the strategy behind them isMelinda Marques and analysts at
Tech targets Enterprise StrategyGroup. Welcome, Melinda.
Melinda Marks (00:46):
Hi, Anton. Great to be here.
Antone Gonsalves (00:48):
Great. Great to have you. So Cisco, how is
Cisco's advancing theintegration between security and
networking in its latest productannouncements from from Cisco
Live last week?
Melinda Marks (01:04):
Yeah, I think they're making a really smart
point about how importantsecurity is to conductivity and
networking. So they made it amajor theme at the conference
last week, you know, Cisco isn'treally known as a security
vendor, or that's not the mainthing that they're known for.
But they made a point of talkingabout how that's the number one
(01:24):
topic for customers. So anytimethere's a connection, it needs
to be secure. Their customersneed to be able to use Cisco
products easily, while havingconfidence that it's secure. So
there are a lot of greatsecurity initiatives that they
announced last week, theyannounced secure access, which
is to make sure that anybodytrying to connect and again,
(01:47):
this is a big point for modernpressures of digital
transformation, you have a lotof remote work, hybrid work. And
they announced that theirsecurity service edge, their
sassy solution really helpssimplify access. So if employees
are connecting from any locationfrom any device, that could do
(02:07):
so and work productively. Andthen on the back end, there's a
easier way to manage thataccess, see where people are
connecting. They also announceda really cool collaboration with
Apple to incorporate zero trustnetwork access for a seamless
native experience for Mac users.So those are really nice, nice
(02:29):
things to see that most peoplecan understand the need to
access and be able to accesstheir work applications
securely. So I think thiscommunications pattern that
they're doing have, make Ciscomore associated with security
was really strong at that event.They also had some great
(02:50):
speakers they had the NFL wasvery present at the show, they
had their party and Allegiantstadium, so they were talking
about how you can have thesegreat experiences securely. So
if you're at a ballpark or astadium, you can go order your
food, you can do everything withyour mobile app, you don't have
to worry about carrying yourwallet. So these are all things
(03:13):
that Cisco's really emphasizingas ways that they can make
modern life better, moreconnected, while making sure
that they're fully thinking ofall the security implications,
so that you don't have to, andyou can just go about what you
need to do.
Antone Gonsalves (03:29):
Right. And they also introduced generative
AI within within their securityportfolio, right? I mean, how's
it going to be used in thecompany's new policy assistant?
Melinda Marks (03:38):
Yeah, and I love that because this is a big
challenge is there's a lot oftedious manual work that's
involved. Cisco made a big pointabout responsible and ethical
usage of AI. There's a lot oftalk right now, with I think,
this year early this year withthe emergence of chat GPT, what
(03:58):
are AI implications, and peoplewant to use it in a smart way to
make things easier and to takeaway manual work and things like
setting policies and settingguardrails. That's something
that takes a lot of manualtedious work. So they're using
their generative AI policyassistant for Cisco secure cloud
(04:21):
so that security and IT adminscan set and maintain security
policies. And this is alsoagain, it's to help them reduce
their manual work. It's to helpmake sure that there's no
duplicated work. These are allthings that really help security
folks, make sure that they canset up policies and easier
scalable way.
Antone Gonsalves (04:42):
Right, and how about generative AI in their
security operations center?That's also an area that they're
going to include the technology.
Melinda Marks (04:51):
Yeah, so they announced leveraging generative
AI for faster threatintelligence, threat detection
and response. So when I'mThere's some issues security
issue, it's really difficult todetect where it is and respond
quickly. So at the end of thisyear and early next year, they
have some new features wherethey have security operations
(05:15):
center assistance, so that youcan contextualize the events and
respond faster, so that you can,you know, get to the problem
faster and, and remediate it.And with those types of tools,
they can better understand whatthe issues are, understand what
their impact is, and make surethat they put in the best
remediation efforts to containthat or stop it as fast as they
(05:39):
can.
Antone Gonsalves (05:40):
Okay, Cisco also launched panopticon, which
brings us into cloud security,right? We're talking about cloud
native applications. I mean,that's, that's new for Cisco.
How will it you know, expand itsits portfolio?
Melinda Marks (05:57):
Yeah. So this was exciting. And I actually got to
speak in their session to talkabout it. So when we think about
the needs for digitaltransformation and moving
workloads to the cloud, there'sorganizations are moving to
cloud native development,because it makes development a
lot faster and more efficientfor developers. And Cisco has
been up to cup and Africa istheir solution for
(06:21):
observability. And visibility,as the name implies, with
panopticon that you can seeeverything. And when you think
about managing security and riskwith the move of workloads to
the cloud, and secure andmanaging security, in general,
in general is always aboutvisibility, and making sure that
you can see into everyconnection, every resource, and
(06:41):
with the dynamic ephemeralnature of the cloud and, and
development resources wherethey're using these. They're
using applications based onmicroservices. So very
distributed, you need thatvisibility to be able to see and
monitor those applications inorder to manage them. And again,
to find security issues and makesure that you can remediate
(07:04):
them. So they have, theyannounced some updates to their
full stack observabilityplatform. And this is integrated
with panopticon. So that's whereyou get that full visibility.
They also have those featuresand capabilities that they had
from AppDynamics and 1000. Eyes.So and they're also using a lot
(07:25):
of their innovation and emergingtechnologies that they found
this as part of their out shiftproject, which is their
incubation engine for emergingtechnologies, including the
cloud, native stack, edgecomputing, AI and quantum
computing. So when you bring allthis together, and you help
people with security, they'rebringing together a lot of
(07:47):
security capabilities that youlook for in a cloud native
application protection platform,which kind of brings everything
together for efficient riskmitigation and application
protection. So now you havethings like, you can monitor the
applications. And you can theyalso have some infrastructures
(08:09):
code scanning capabilities,workload monitoring
capabilities, and applicationand API security capabilities
that they demoed in the sessionthat I was in that show how this
gives people a platform thatties in what what assets they
have in the cloud, how tomonitor the workloads, and then
identify any issues and makesure that they have fast,
(08:31):
effective feedback loops fortheir developers. So when an
issue is found, it canimmediately go back to the
developer, so the developer canimmediate it with speed and
efficiency. So we
Antone Gonsalves (08:42):
look at Cisco and Cisco's portfolio is huge.
We know that we could spend twoand a half hours just going
through all the all the featuresthat it provides. But you
mentioned early on about Cisconot being seen as a security
vendor security provider, atleast not a leading one.
Melinda Marks (09:03):
Networking, yeah, use extensive
Antone Gonsalves (09:06):
portfolio, you would think the opposite? Is it
because Cisco has seen itssecurity portfolio was seen as
something closely tied to itsnetworking and therefore
something most beneficial toCisco shots, enterprises that
are heavily invested in Cisco.And
Melinda Marks (09:24):
yeah, absolutely. And I think this is why I'm
excited that organizations arelike Cisco, where security is so
important that they meet thishead on right. I've been at
technol. I've been on the otherside at technology vendors,
where security is an issue. It'syou know, when you think of why
would people not buy my productsecurity, if it can be a
(09:47):
blocker, they think it'sinsecure. You know, pricing and
other things might be a reasonwhy people might be wondering
what whether they should buy itor if they should choose another
product and I I really like thisboldness of Cisco to just take
it head on and say, No, we'regoing to we're going to meet
these needs and where you wantto be a major security vendor,
(10:10):
we want Cisco Live to be a topsecurity show that was kind of
surprising to me and all theirpress releases. This is a
networking and securityconference, where they're trying
to position it that way. So, youknow, to me as a security person
that thinks that's the mostimportant thing that people
should care about. And it shouldbe an enabler and not a blocker.
It's just really refreshing tosee them put that emphasis. And
(10:31):
then also, because they havethat reach of the types of
products that they have, likethey're known for their network
security, they're sassy, theydid announced some firewall
updates as well. You know, thisreally expands into like the
area that I cover, which iscloud and application security,
and helping with digitaltransformation. And they do have
(10:52):
a lot of advantages therebecause they have the different
components that they need thatyou need to address that well.
So things like securing at theedge for the developers as
they're using their laptops towrite code spanning to the
network spanning all these areasthat you really need to do and
get control of and disabilityhave an order to mitigate risk.
Antone Gonsalves (11:16):
While US lawmakers did or over AI
regulations the European Unionis made is moving toward the
passage of an AI act. Here totell us the impact of the EU's
aggressive action on AIregulation is techtarget.
Tutorial government reporterMacKenzie Hollen. Welcome,
Mackenzie.
Unknown (11:36):
Thank you, Anton happy to be here.
Antone Gonsalves (11:37):
Right. So what's in the AI act? And where
is it on the road to approval?
Makenzie Holland (11:42):
The AI act is legislation that essentially
classifies risk of artificialintelligence and high low and
prohibited artificialintelligence use. So
essentially, lawmakers havedeemed that, you know,
prohibited uses of artificialintelligence include facial
recognition, or using AI toclassify people based on
(12:05):
socioeconomic status or otherpersonal characteristics. So I
think the most significant partof that legislation is the way
that a classifies AI into risk.So essentially, you have
different obligations to complywith based on the level of risk
that your AI is classified at.And so where they're at in this
process is the EuropeanParliament just approved this
(12:27):
version, this draft version ofthe AI Act. Now it goes to the
European Council, and they'regoing to be speaking with other
member countries to finalizethis law and essentially propose
what the final version is goingto look like. And it's already
moving a lot faster thisprocess, then people thought
that it was going to so this isalready expected to be a final
(12:47):
loss expected to be proposed bythe end of the year.
Antone Gonsalves (12:51):
Okay, and what's the what's the impact
potential impact on UScompanies? I mean, what are your
sources telling you?
Makenzie Holland (12:59):
Sure. So it's, there's going to be obviously a
time period where companies aregoing to have you probably two
to three years to comply withthis legislation. So there won't
be immediate impacts. But it'sgoing to be something that, you
know, companies will have toadjust to, because they even
have a section in there forgenerative AI. And that's
(13:20):
something that we've been seeingtaken off recently. So
essentially, companies are goingto have to be transparent about
their generative AI use andgenerative AI is going to have
to essentially tell users ifthis content was aI generated,
and there are other transparencymeasures behind that. So
companies are going to have alittle bit of time, you know, to
comply with this legislation,obviously, once it is approved
(13:43):
and finalized. So it's hard tosay what the immediate impacts
would be, but it's there. Peopleare gonna have some time.
Antone Gonsalves (13:50):
Yeah. When you talk about transparency and
their generative AI us, doesthat mean, they got to hand over
their algorithms to the EuropeanUnion or something, something,
Makenzie Holland (14:00):
you know, it's, again, these this is not a
final form of this law. And it'sa kind of broad section. So I
think it's up for interpretationa little bit. But yeah, right
now, it's just essentially kindof asking for, to tell whether
or not this content is AIgenerated. Have the training
data on hands, essentially, kindof the I don't think they're
(14:23):
asking to turn over thealgorithm itself, but just for
transparency into outputs. Andwe'll see where it goes from
there.
Antone Gonsalves (14:33):
And in the US, so either what are some of the
proposals being brought forthand as to how we would regulate
AI if we ever decide to
Makenzie Holland (14:45):
regulate? Yeah, there are a lot of
different thoughts on that.Following open AI CEO Sam
Altman's testimony back in May,Senator Michael Bennet proposed
a bill to create a federalagency as the typically to
regulate artificialintelligence. That's one thing
that's been proposed. Altmanhimself proposed, you know,
(15:07):
licensing requirements for largeAI models. And then even
yesterday, Senator Josh Hawleyintroduced a bill that would
carve out generative AI fromsection 230. immunity, which
section 230 protects platformsfrom content generated by third
parties. And so they'retargeting AI that way.
Antone Gonsalves (15:27):
And there's also me there are people who
favor the tech industry,regulating itself set their own
standards. I mean, that sounds,that doesn't sound like a great
solution to me, because you are,you're putting the fox in charge
of the henhouse and AI isimpacts so much of our lives. So
(15:50):
it has the potential ofimpacting so much of our lives.
So what are what are the prosand cons of that particular
approach?
Makenzie Holland (15:57):
Sure. So I guess I could touch on the cons
of that first, becauseobviously, we've seen what's
happened with companies and selfregulation, when it comes to
social media platforms.Obviously, there wasn't a lot of
self regulation going on when itcame to data collection,
privacy. And so we've kind ofbeen seeing some issues with
(16:17):
social media as a result of kindof letting companies and do
their own thing with that. Someof the pros, these industry led
standards have underlyingtechnologies like Bluetooth like
Wi Fi, and, you know, there'skind of a universal agreement on
how these kind of tools shouldbe used. So it's, you know,
(16:37):
there are pros and cons to theapproach. I think it's just
going to depend I don't thinkwith artificial intelligence is
going to be left solely up tothe companies there's too much
and what we're seeing here,Europe already with the AI act,
they're not going to just letcompanies do this on their own.
Antone Gonsalves (16:54):
Right. So it seems like the tech industry is
going to have to, and theyalready are worked very closely
with government. And, andhopefully consumer groups to be
involved to to protect, protectthe rest of us. This week,
Oracle and Salesforce made bigmoves in generative AI Oracle
(17:17):
said it would offer its cloudcustomers, startup cohere for
building generative AIapplications, Salesforce has a
different approach. Thecompany's new API's Cloud Suite
includes an Einstein GPT trustlayer that prevents generative
AI models from holding yourcompany's customer data. Here to
(17:38):
discuss the latest moves fromOracle and Salesforce is tech
targeted tutorials AI reportingthe rest of his job. So what
does cohere offer companies thatwant to run AI applications on
Oracle Cloud,
Esther Ajao (17:52):
so are going for air are kind of working together
so that customers can traintheir own customized large
language model. So what articleas well as a melee other cloud
providers but melee are goodselling point. And also
Salesforce says we're going todiscuss later on is this idea of
protecting customers data. Andso what they're offering
(18:16):
together on OCI is that theywill plan on offering like a
service that will not mixcustomers data with Arcos data
or with chorus data. So coreselling point is they are very,
like, your data stays your data.And we are not going to mix that
together, if that makes sense.
Antone Gonsalves (18:35):
So the I take it customers would use their own
data to train the models isYeah,
Esther Ajao (18:43):
yeah, customers will use their own data to train
the model. But with like others,right, you especially the
problem, not the problem. Butmany of the challenges or some
of the drawbacks that customersenterprises have with tools like
open AI chat. GPT is this mixingof data, so to speak, their
concern that perhaps likesensitive data, especially if
(19:05):
you're working in like, thosesensitive fields, like the
healthcare industry, will kindof lead in and you will be their
data will be used to train thesystem. And so cohere,
specializes in not doing that.And so this partnership is kind
of good. It's very good forOracle. So it's good for us. So
Correct.
Antone Gonsalves (19:23):
Okay, what tell us a little bit, a little
bit about the status quo hereand its relationship with with
Oracle.
Esther Ajao (19:30):
Yeah, so Korea is kind of like open AI. But they
work primarily with enterprises.And so they let enterprises use
their own data to train the AImodels. So they're not sharing
the data, the data stays, itstays within the customer stays
within the enterprise. And solast week, they announced that
they have raised 270 million,and part of the fundraisers are
(19:54):
part of the people involved inthat fundraising is Claire, I'm
sorry, it's Oracle as well asNvidia And so Oracle, obviously,
is investing a lot of money intoCO heir as other companies. But
this is a good relationship forOracle, who can't like
necessarily go to my open AI,which already has a partnership
(20:14):
with Microsoft, as you very wellknow.
Antone Gonsalves (20:16):
Yes, I believe Salesforce ventures A, which is
competitors also invested in COhere.
Esther Ajao (20:24):
Yeah, yeah, hearing voices us also invested. So it's
not like coin is like married toProko. They have the investor,
as well as they also have a lotof they also like, I think
they'll also have products thatcool air is like, that
integrates chorus technology.But this is kind of like a
(20:45):
little bit of helping Oracle getinto the large language model,
generative AI market. Obviously,Oracle has a big database, that
will be helpful to Korea aswell.
Antone Gonsalves (20:57):
No, I mean, it would seem that the the approach
that cohere takes, who iscertainly in line with what,
what enterprises want, I mean,I've talked to people who say
that they won't even usegenerative AI, unless the vendor
can show can prove that the datais, is protected.
Esther Ajao (21:21):
Yeah, that's, that's, that's, that is quite
shallow, but that I thinkthere's a lot of things that
enterprises are like, concernedabout is that data aspects,
obviously, like I mentioned, ifyou are in the sensitive field,
you don't want your data to beused to train a model like I
don't want, I don't want my, youknow, like, I don't know, like
my DNA sample, or like myhealthcare information on chat
(21:44):
TBTF, so to speak, like that.And so that was I think, even
just to re emphasize was, thatwas one of the selling point of
their whole Einstein, GBT trustlayer is that you will have that
trusted asset, that trust thelayer that you really need, in
order to so that you can you canknow that your data is secure,
you can know that your customerdata is secure. And it's not
(22:07):
being used in this AI system.And also it's kind of
explainable, you know, wherethat data is coming from, you
know, it's not going into some,some black box or you don't know
anything about
Antone Gonsalves (22:17):
That's right. But Salesforce, Salesforce
Salesforce is approach, I usedto be a little bit different,
because you're not actuallycreating the AI application.
They're gonna give it to you.But they have that control layer
that the data from from fromheading to someone else's
(22:37):
database.
Esther Ajao (22:40):
Oracle and curry I taken their approach of you will
be able to make their own LMmodels, you will have new app
access and controls everythingout to everything. Obviously
Oracle also plans on embeddinglike core technology and models
into their businessapplications. But it also gives
the control to the enterprise aswell as Salesforce is like we
(23:01):
will have that extra layer thatyou need to know that first show
this is really secure. Sodefinitely two different
approaches.
Antone Gonsalves (23:09):
Okay, and when is is coherer available now on
the Oracle platform and whatabout that Einstein GPT trust
layer? Is that available now orin the future?
Esther Ajao (23:22):
I believe it is not available. I believe it is in
the future. I'm not I don'tremember that on
Antone Gonsalves (23:30):
All right, so that wraps up this week's show.
Thanks for watching. Enjoy theweekend.
Popular Podcasts
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com