KCAA: Inside Analysis with Eric Kavanagh (Sun, 16 Jun, 2024)

KCAA: Inside Analysis with Eric Kavanagh on Sun, 16 Jun, 2024

Episode Transcript

(00:00):
Dot Comsideanalysis dot com, and nowhere's your host to Eric Kavanaugh and Flare.
All right, ladies and gentlemen,Hello and welcome back once again on
the only coast to coast radio showin the US of ATA. It's all
about the information economy Inside Analysis.Yours truly, Eric Kavanaugh here, and

(00:20):
folks, I'm very excited to bedialed into our nation's capital. We're going
to talk to an industry visionary today. We've got mister Chris Moore on the
line from the Software and Information IndustryAssociation or SIIA. We're going to talk
all about data and the data lifecycle and protecting the data life cycle and

(00:41):
frankly being responsible stewards of data.And that's good for lots of different reasons,
one of which happens to be thisexplosion of artificial intelligence. So it's
not just people using data and peopleprocessing data, and now there are machines
that are using data to train onand to do all sorts of different things.
And that is very, very disruptive, and I think it's going to

(01:02):
continue to be disruptive. So wewant to talk also about responsible AI and
how to get there and the shortanswers a lot of it has to do
with data. And with that,let me welcome Chris More to the show.
Thanks for your time today, tellus a bit about yourself and about
the SIIA. Sure, my pleasure. It's good to be with you,
and thank you for the invite.Thank you, Ben. We are What

(01:26):
we like to say is we're aplace for folks who are in the business
of information. And what we dois we look to preserve and protect a
healthy information life cycle. In otherwords, we want to make sure that
there is a good environment for itscreation, dissemination, and productive use.

(01:46):
And our members range from platform companiesyou're probably well familiar with, as well
as educational publishers in news organizations anddatabase publishers and financial data firms. Okay,
and you get into a lot ofdifferent policy areas obviously. I mean

(02:07):
there's intellectual property, there's privacy,there's cybersecurity, and just the use of
data. So when you talk aboutthe data life cycle, I presume you're
talking about data from its creation toits grave, basically, from its inception
through its use all the way toits deprecation, which in many cases doesn't

(02:28):
even happen because it just sits aroundforever. But a lot of cases,
companies will follow process and procedure andafter X period of years, whatever the
regulation is, they will delete thatdata and move on. Can you talk
about how that works? I mean, you are an association, so you
have leaders from many of these bigcompanies I guess over eight hundred members who

(02:49):
come together to forums and really talkabout this stuff and kind of hash out
what to do. Because I knowfrom experience policy is hard, and policies
need to be understandable, they needto be known, and they need to
be reasonable in order for people tofollow them. So tell us a bit
about how you go about doing thatand maybe give us some examples. Uh.
Sure, I mean I think therethere are areas, for example,

(03:15):
uh, where there's a pretty goodbit of consensus within our membership, which
is the easiest place to take policyrights. And so for example, let's
talk about let's talk about privacy.That's right, that's a hot topic.

(03:37):
And is there an AI wrinkle?Absolutely there is. Uh. But that's
an area where we saw a bunchof you know, we saw state laws
coming from California in their in theirsignature law, and uh we looked at
the text and we saw, okay, you know, there's a looking at

(03:59):
it from the standpoint of a UnitedStates model, which it was likely to
be initially, this has a realFirst Amendment problem. And the way that
Europe approached GDPR and the way thatwe as the United States approach the use
of information are very, very different. And so there we looked at that

(04:23):
and we said, look, wewant there to be a privacy law.
We're not opposing the idea that youpass a privacy law, but you need
to carve out publicly available information becauseif you don't, the whole thing will
go down in claims, right,And so we successfully on behalf of our
members and they were uniformly, Ithink behind this advocated for language in that

(04:47):
first in the CCPA and then inthe ballot initiative that carves out the use
of publicly available data. So youcan't tell somebody who publishes, as a
for example, of database of newsarticles to stop selling information about you or
to delete information about you, whichunder GDPR is you know, theoretically it

(05:11):
can have and certainly in the businessto business publishing space, which is another
division of ours, that that kindof thing is really important. Yeah.
Well, and you know, Iremember when GDP R came out and they
had this whole thing. I thinkyou might like this. They had this
whole thing about the right to beforgotten, which is where they're saying that
if you're an EU citizen and youwant some corporation to delete whatever data it

(05:35):
has about you, that you havea right to do that, and the
corporation has to follow through and goahead and delete your data. Well,
as soon as I heard that,I was like, good luck with that
one. First of all, dataabout someone in a company gets lots of
places. It could be in thisdatabase, it could be in that database,
it could be in email exchanges betweenpeople. There are lots of places

(05:55):
where that data can live, andquite frankly, in many organizations, there
aren't too many people who know allthe places where that data can live,
and there it's hard to query yourentire information landscape. I mean these days.
Just this morning I did a callwith Justin Borgman, who's the CEO
and co founder of a company calledStarburst, and they do federated queries like

(06:15):
on a data lake, basically wherethey can query all sorts of different systems,
all sorts of different databases through oneabstraction layer and that's very cool,
and it's also very new, right, So, and not every company has
Starburst or Dramio or these kind oftools. So point being, it puts
a tremendous onus on the organizations tobe able to do all that. Besides

(06:36):
which it's very difficult to know.Is this the John Brown who lives at
one on one Main Street of theJohn Brown who lives at four o one
First Street or something like? Justknowing identity resolution is a challenge in and
of itself. So my point isGDPR pretty owner is stuff. I mean,
I understand it's not always prosecuted.So the idea is to have these
sort of goals to strive toward.But there's something I want to throw out

(07:00):
and maybe just get your take onall this stuff. I came up with
this concept. I call it theright to be respected. And what that
means is I want anyone who's usingmy data to respect my data and to
respect my preferences. So if youask me how do I want to be
contacted, I say by email,not by phone. I would like for
your operational system to reflect that andfor people to not call me and said
to email me. Right, thatrequires an architecture with policies that are baked

(07:26):
in and adhered to. But Ithink that's a more reasonable approach is to
say to the X y Z organizations, yes, of course you need my
data when I buy things or ifI sign up for stuff. That's all
fine, but just you know,be a careful and responsible steward of information
and have some policy around that thatmakes sense. But what do you think

(07:46):
about all that? I mean tome, that sounds very sensible, and
I think you know the question thatwhere it becomes where it becomes sticky,
I think is depending on who youtalk to. The idea the word respected

(08:07):
carries a lot of weight, rightright, That does a lot of work.
And so you know, for somepeople respecting data men, you know,
the second, the second that Ihave conduct finished conducting business with your
organization, I want you to takeall my data, put it on a
hard drive, smash it into bits, and throw it in the sea so

(08:28):
I know that it's gone. Youknow, that's probably not terribly practical,
right right. And other folks maysay, no, that's not generally,
not how our well not generally,it's not how our members roll. But
they won't say, well, werespected it we respected you enough to get
the data from you by consent inthe first place, Right, are we

(08:50):
done here? That's not Yeah,that's that's not that's not practical either,
and it's Look, nobody wants toget spammed with tons of messages from a
company that they just did business,would want or God forbid called, right,
And I think are there are somenatural market checks on that stuff.

(09:15):
And that's an excellent point, right, is that the companies that do respect
your privacy, that do respect yourdata and you as a person, well,
they're going to have a good brand, they're going to have a good
reputation. And that's where you letthe market guide things. But to your
point, you do need some kindof policy, and I think the good
news is that the technologies we havetoday enable robust policy management. Ten years

(09:39):
ago or twenty years ago, allyou can really do is govern access at
the database level or at the applicationlayer, for example through a log in
or something like that. But nowyou really can't have policies that are baked
into cloud based systems that are usuallyrole based. If this person is an
accountant, they have access to accountingdata. If they're a and your executive,

(10:00):
the access to a lot of differentdata. All those kind of things
are possible today, and that thatmakes the process of policy design and implementation
and enforcement a whole heck of alot easier, right, Oh for sure,
I mean, I think, andthat's something even as a small organization,
you know, we have where thereare people, you know, there

(10:22):
are people who need to have thekeys to everything, but there ain't there
ain't many of them, right right, right, And there are there are
folks who need to be able toaccess Microsoft Office. And that's essentially about
it, right Uh, And it'sreally important internally. I think technologies come

(10:43):
a long way to allowing internal companiesto exercise those controls as well as I
mean in terms of what our membersdo. They even when they're dealing with
customers, they there will be arange of controls on what a particular customer
might be able to do with whatthey access and how that's determined. So

(11:05):
that could be everything from refs andwarranties in a contract in terms in terms
of use in the subscription agreement,to IP monitoring and clearance, to sometimes
even site visits to be sure thatthis place that is licensing access to information
is you know, in fact exists, right, and it's not some you

(11:30):
know, it depends, it's theyadopt a risk based approach to all of
this stuff. Yeah, that's agood point. And you make another good
point about having the need for peoplewho understand how the systems work. Right,
if something is a black box,well it's kind of hard to understand
a black box. And now,at the risk of getting ahead of ourselves

(11:52):
a little bit here, artificial intelligencelarge language models, right, these companies
that build some of the first ones, Open AI, for example, I
know a lot of people who knowabout that whole process. And what I'm
told is that early in the gamethey realize they cannot just set these engines
loose on the Internet at large,because there's a lot of nonsense on the

(12:13):
Internet. There is a lot oftrue stuff, and there's a lot of
stuff it's not true. There's speculation, there's sarcasm, they're all kinds of
things, and so they had tobe very careful about how they trained the
models. But they clearly train themodels on a corpus of data on the
Internet. And so where do yourun into how do you even delineate copyright

(12:33):
on something like that? And what'sfun is. If you ever ask chat
GPT or Gemini, which is whatI use about copyright, it will give
you a pretty thoughtful answer. I'llsay, well, actually, you know,
you might want to talk to yourattorney or something like that, but
what are your high level thoughts aboutthat? I realized it's a very sensitive
subject, but I'm sure you guysare talking about that, right we do.
And this is one This is anarea I had mentioned. There are

(12:56):
areas where members are agree, andthis is an area where they don't.
But there are there are a fewpoints. And I'm you know, I'm
a recovering copyright lawyer. Uh andso you know, perfectly happy to have
a relapse. But just so youknow, uh, you've been there,

(13:20):
I've been there. So there arethere are definitely two sides to this story.
And the on the one side,you have a technology that at its
core is looking for statistical relationships betweenwords and series of words. Basically,

(13:46):
right, that's what it's interested in. Right, So a word, you
know, quick brown, the nextlikely word is fox is a lot more
likely than ruda, bega. Andthat's it's a physics model of sorts for
language, and so in determining thoserelationships. You are abstracting out a particular

(14:07):
non fact, if you like.And the argument on the one side would
point to reverse engineering cases that haveheld that type of life, that type
of activity to be lawful. Right. That's one piece of it. The
other piece of it is that onthe one hand, the work may be

(14:31):
copied. The work was copied,there's no doubt about that. It is
being used to create works that maynot be literally similar but are nonetheless competitively
substitutive for the works of the originaloriginating authors. That's an issue, and

(14:52):
there may be issues. There maybe other issues in there around the weights
assigned to particular kinds of content.There may be issues around whether or not
any of the content was parradical,There may be issues around any number of
things. But the thing is isthat these are issues that are uniquely poorly

(15:13):
situated to have a legislature deal withthem, because they're fact specific, and
that's how the courts resolve these things, one case at a time, and
that's where our members are. Soeven though they disagree about these things,
and they think in some circumstances thelaw to play one group will say,

(15:35):
well, in this circumstance, thelaw to apply in a particular way versus
a different circumstance. That kind ofsalami slicing is what needs to happen before
there can really be any policy solution. That's very interesting. So for example,
I remember, well, I guessin Canada they just hit passed this
law. In California, I thinkthey're doing something similar where the government is

(15:58):
saying that if you link to abunch of news articles from media companies,
you have to pay those people basedon the clicks and things of this nature.
And so you know, what's thelaw of undertended consequences doing here.
It's saying, well, finally,we just won't do that, so we
won't disseminate the news anymore. Wewon't disseminate this information. And it gets
it kind of gets sticky, veryvery quickly, and you start asking yourself,

(16:21):
all right, does this make sense? And like, how would you
compensate people for these things? Wouldyou have to open the kimono and show
what you paid as an organization tothe author? I mean, wow,
that conversation gets goes sideways pretty quickly. And I remember news aggregation sites,
for example, where you just havea bunch of links to other articles they're
saying, oh, now you'd haveto pay those people. This has coming

(16:42):
gone a few times. We've gotour first break coming up here in just
a second, but picked this upafter the break. It is a bit
of a stifling question to wonder,well, hyperlinks are what the Internet is
all about, Like at the verycore, it's just links to other stuff
that you can then excit floor andlearn about and and have discovery and do

(17:03):
all these kinds of educational things.So if you start trying to track that,
how would you even do that?I mean, there's no one standard
for being able to report on thatis everyone have to download a certain kind
of software to that. I mean, it just it goes sideways so quickly,
and I don't think that many peoplereally understand that a lot of these
aspirations are either untenable or going tohave unintended consequences that are just so dire,

(17:27):
and it really undercuts the whole missionin the first place, of trying
to protect the authors of some particularpiece of content. Somewhere it gets wild
and wely real fast. But folksdon't touch that. Dell will be right
back with Chris Moore. You werelistening to Inside Analysis. Welcome back to

(17:52):
Inside Analysis. Here's your host,Eric Tavanaugh. Right, folks, take
us to the future, indeed,talking all things information, information, life
cycle, responsible information management. Sowe're talking to Chris Moore of the Software
and Information and Industry Association, Rightthat right, then does it? So?

(18:21):
S I I a dialing in fromour nation's capital, Washington, d
C. And you mentioned in thelast segment for radical meaning pirateed and so
if someone has trained their model onpirated software, pirated movies for example,
that's a whole separate area that needsto be handled differently from a policy perspective.
Right, do you want to talkabout that for a second. Sure,

(18:42):
I mean I think I'm talking aboutwhen I talk about that, what
I'm talking about is in the contextof a copyright lawsuit, which is where
all this is now. There isuh there's there's a defense called fair use,
which is kind of copyrights. Ifyou like copyrights, golden rule.

(19:03):
It's not due unto others and runit's it's to others as you would have
them do uh do to you.And it's designed to be a rule of
reason, right, and so,but it's one that's based in equity or
fairness, and so your risk looksdifferent if you had kind of lawfully acquired

(19:27):
a bunch of stuff, a bunchof legitimate copies, if you legitimately got
access to legitimate copies and use them, versus you've got access to illegitimate copies,
infringing copies and use those, thefairness balance looks different. It's the
risk. The risk analysis is simplydifferent, you know. And that's saying

(19:48):
it without opining on how the casemay turn out, right, because you
there's lots of other facts that couldswing it one way or the other.
But the risk looks different, right. Yeah, that's a very good point.
And so what we're trying to dois to find reasonable policies and I
know all about fair use. Itbasically says, look, I mean you

(20:11):
can read something and remember it,and then if you happen to quote it
and you didn't realize you're quoting itverbatim, well you're not doing something bad.
You're just reflecting back. And what'sinteresting is that's really kind of what
these AI models are doing, thelarge language models. The big AHA moment
for me was understanding that when youtrain one of these models on information,

(20:32):
it's not actually persisting the verbatim text. Rather, it is adjusting. The
engine adjusts its own weights and biasesand its parameters, which is kind of
like how humans learn. I mean, these are neural nets that were designed
to attempt to reflect how human beingsbehave. But the point is it's not
like copying and pasting later on.It's absorbing this information, adjusting its parameters

(20:57):
accordingly. And there are billions ofAMers in some of these models. And
then it's reflecting back to you textbased upon your prompt. And it's a
predictive engine. Basically, it's apredictive engine that gives you what it thinks
you're looking for based upon your prompt, which is a very different thing from
just copying text and then pasting itunder a different name. Right, yeah,

(21:18):
I mean, look, I hada similar similar insight for me was
that these engines are essentially a particularkind of artists, which I wouldn't describe
on a family show. I wouldsay. What I mean by that is
they say, I mean, youknow, act as if you were so

(21:41):
if you ask them a question andsay, you know, I would like
the answer to this to a particularquestion, can you explain the fair use
defense. To me, you weresaying, act as if you are an
expert on fair use, what wouldyou tell this person right? And then
the quality of the data will determinethat answer. Right, it could be

(22:02):
right, it could be wrong.And when you get into more sensitive questions
around security, you know, aroundfor example, I don't know, creating
chemical weapons and stuff like that,it requires a lot of thought and care
to make sure that the engine doesn'tdo what it's supposed to do. So

(22:25):
to speak right well, And ofcourse there's dual use technologies. That's very
common thing to watch out for inthe government, because you can use the
stuff for fertilizer or to make abomb, right, And so you have
to be able to understand the differencebetween the two. And you know you're
right that this comes into question whererubber meets row when there's a lawsuit somewhere,

(22:48):
and then you have to hash outall the different issues and try to
understand. But I think you're you'reon target to say that the fair use
doctrine is meant to be fair,as the name would indicate, it's meant
to say, all right, thereare lines we can draw about how you
can responsibly use these things, andI understand it. In the music industry,
there's some pretty precise rules around that, like how many notes in a

(23:11):
row can be the same, andthings of this nature when push comes to
shove in a court of laws somewhere. But just generally speaking, it's very
interesting to me to see how theselarge language models have triggered all these conversations
around ethics and ethical AI and whatyou can do. And I think the
answer to most of that is solvedin two different ways. One is data
governance and the other is transparency.And the transparency argument. I'd be curious

(23:36):
to hear your thoughts on that,because of course you got LAMA two and
LAMA three from Meta, which areopen source. Open AI used to be
open source, it's not now it'snot a black box. And what do
you think from your perspective, howbig a difference is that? How does
that change things when it's a blackbox versus an open source model? So
that is I'm going to give youan answer that's accurate and both technically accurate

(24:03):
and practically useless. The answer isthat it depends right. I think you
know because we've seen we have seenthese debates before, and by that I
mean, you know, there wasa time when those who made and i'll
put them in scare quote, putit in scare quotes black Box software,

(24:26):
were scared to death of open source, viewed it as a virus, one
that needed, you know, aproportionate response, right, and it turned
out to be a really good wayof developing secure and interoperable systems in certain
circumstances. There are times when youwill want that, there are times when

(24:49):
you won't. I think here weare seeing to some degree, a similar
kind of we're seeing a similar dramaplay out. I think it is a
little different around the edges, butthere it really has to go. The

(25:11):
way that it's different is that theengines themselves are so flexible and so there.
In other words, if you build, if you're building I don't know,
you know, an open source videorendering tool just looking at this screen,
that's what it is. Whereas anopen AI, an open sourced AI

(25:33):
model is different. So what isthe right approach to that The right approach
to dealing with something like that isgoing to be based on risk. In
other words, you have to besure that as and this is the way
that the US is progressing, thisis the way the White House is made
clear that it wants to go.They want a risk based approach to this.

(25:56):
To give you a really simple answer, in our experience, I mean,
we've been around for going on we'recreeping on fifty, which is a
long time for a quote technology closequote association. So you know, we've
seen a few rodeos, and inour experience anyway, technology neutral regulation tends

(26:23):
to be the best way to approachthese things. In other words, you
don't look so much. You lookat the risks of the technology creates.
So fraud is still going to befraud. It doesn't matter if it's a
scammer calling you in real life,where if it's a bot scammer using AI
to create a voice. They're bothokay, both of those things are scary.

(26:45):
They are both. It's a dualuse technology, right. They will
use it in movies, for example, to create a character that no longer
exists or an actor's voice that canno longer use his right. Right,
that's dual use technology. But it'sstill fraud, right, and there are

(27:07):
remedies for that, And if thereare technology specific angles, that's what we
need to look into as opposed tosaying you know, okay, and this
is typical of the EU sort ofwhere they started was it's not coming out
unless we approve it, right,and that's that's treats you know, your

(27:27):
Spotify algorithm the same way that itwould treat you know, power grid management.
Mm hmm. That's just not thatdoesn't make sense. Yeah, and
you're reminded me of something too,because deep fhase are definitely an issue.
Certainly we're going to see this inthe political sphere. I've seen a couple

(27:47):
that were very funny. I'm notgonna lie. I won't say which one
it was or who it was,but it was like, what what did
she say? We had to laughbut were Okay, it's a deep fake,
that's what's happening here. But Ithink that probably and i'd be curious
to your thought on this. Thereare some very practical ways you can leverage
technology to know if something is genuinebecause most cameras are digital cameras these days,

(28:08):
and they have a whole host ofmetadata that's baked into that device about
you know, what device it is, what time this thing was taking,
what the location was, a lotof them have geolocations, so if you
pass all those markers, then thatlooks good. Then you could say this
photograph was of the wreckage from atornado in Kansas, and it checks out.
Yep, it was in Kansas.Yep, it was this date.

(28:29):
Okay, that looks pretty good.And you're starting to see I think it
was meta and someone else say maybeit was Google. They're going to try
to create some watermarks on digitally alteredphotography videos things of this nature. Also,
I think that's I think that's avery practical way to go, because
then you'll be able to sense thator your browser can sense that, and

(28:52):
it'll say and you can turn iton or off, like is this real
or is this generated? I thinkthat's some pretty good stuff. Even when
someone calls you on the phone,is it a real voice? I mean,
you know, I get calls andI know the technology enerity, so
I'm curious to see and like they'llwait a second, Oh, how are
you doing anyway? This is Janefrom such and such Up up, I'm
like, okay, sure it is, like this is a computer voice talkingy

(29:15):
But what do you think about allthat? About the practical ways we can
leverage technology and existing formats and protocolsto kind of shepherd ourselves towards getting to
the truth. So there's one ofthe ideas I tend to be optimistic about
the role of technology in this space. In other words, we will deploy

(29:38):
things and then there will be therewill always be some kind of unintended consequence
right right one way or another,and then we figure out a way to
fix it. So in this particularspace, content provenance, what you were
talking about is that's exactly the that'sexactly the kind of thing that can greatly

(30:03):
ameliorate the harm from from defix becauseyou'll know where the image came from and
whether or not it's faltered, right. That's you know, that that kind
of thing. It takes time.Standards take doesn't take long to lie right,
but it does take time to developstandards for figuring out what the truth

(30:26):
is, particularly when you have youhave a variety of companies with a variety
of different kinds of IP getting togetherand trying to figure out something that's going
to work for everybody. That's howthe standards process works, and it's an
incredibly valuable one, but it takestime right, right, And we've seen
we've seen the same kinds of thingshappen in all kinds of areas, you

(30:48):
know, content moderation, I know, it's a it's what they call a
fraught topic, right, right,But those algorithms get a lot of bad
stuff down, They get a lotand yes, is there are people doing
bad things? Yes, are theydoing bad things on the internet? For
sure? And it takes a andit doesn't take the scale is so the

(31:14):
numbers are just so big that evena small failure rate is an enormous amount
of available bad stuff. And it'sand it's hard to you know, it's
hard to persuade people that they're peopleare trying really hard when the numbers are

(31:36):
so big. But it's you know, it's kind of like from the oldest
proving your value as a council throughall of the bad things that didn't happen.
Yeah, that's funny, right,Like you can't you can't prove a
negative, right the age old argumentor if we if they did, things

(31:56):
would get really really bad. Right. Well, there you go with the
law of the law of unintended consequences. And you know, boyd, can
it come just screeching down the pikeif you throw down some mantra or some
dictate about what can and cannot bedone. Again, I get to what
I call the black market effect.You know, when when rules and regulations

(32:17):
are unreasonable and significantly unreasonable, youcan rest assured that people are going to
go around them. They're gonna findall the ways to do it. They're
gonna offshore it, they're gonna getdeputies to do it, they're gonna pay
people other ways to you know,have them do it. I mean,
there's just like a thousand things youcan do, and getting to the bottom
of it is a very difficult thing. You know. I had a guy
on this show long long time ago, I think in two thousand and nine,

(32:38):
a guy I never forget his name, Chuck Nice from the Insurance Institutes
of America, and he had thisgreat quote. He said, let me
tell you, if an insurance companywants to hide something, you're not gonna
find it. It's like, wow, you know, I hadn't thought of
it that way. He had agreat quote too, because we were talking
about it. Could better governance haveprevented the housing in two thousand and eight

(33:00):
of the market crash, basically themoney markets that went down and all that
stuff. And he said, yeah, you know, I don't think so
he goes, who would have knownthat lending money to people who couldn't afford
it was a bad idea? Youknow? So you get back to that
practicality argument, right. I mean, you can talk all day about theories
and policies and all these kinds ofthings, but at the end of the
day, it has to be reasonable, it has to make sense. Well,

(33:22):
folks, second segment is up here. Stand by, we'll be right
back. You were listening to InsideAnalysis. Welcome back to Inside Analysis.
Here's your host, Eric Tabanac.All right, folks, back here talking

(33:43):
to Chris Moore with the Software andInformation Industry Association or sii A. And
Chris, here's a topic that I'mdying to learn about data brokers. There
is so much data being bought andsold. There are all these contracts out
there, A lot of good stuffis happening. A lot of people have
no idea that this is happening orwhat's going on with it. But give

(34:04):
me your take on data brokers andwhat the policymakers are trying to achieve in
that whole world. So I think, well, what they're trying to achieve
I think is the question is notwhat they're trying to achieve, because we
share those goals, is really aquestion of method, and they're very much
at times it feels like they're verymuch at the world Peace stage of thinking

(34:30):
the World Peace statement. The problem. The problem we have is that the
way they typically define data broker isthat a data broker is somebody that publishes
personal information about someone else with whomthey don't have a direct relationship. That

(34:50):
is incredibly broad, and it isOur point of view is if you look
at that's also a kind of ifyou think you can just come in and
regulate that, that's kind of aEuropean way of looking at it. In
other words, that this property,that this information is the property of the
person who you're talking about. Thatif you look at things that way,

(35:17):
then the data broker regime makes sense. The problem with that is if you
are talking about an entity that collectsinformation about other people with whom they don't
have a direct relationship, what you'retalking about is regulating the activity of commercial
publishing, and that has and youcan do that, but you have to

(35:43):
look at it through a different lens, and there's a whole bunch of reasons
why personally, it doesn't make senseto think of personal information as property.
And I'll give you a couple.So the first reason is we have a
first Amendment. So there is onetranche of information that you can't regulate at
all, right, and that's stuffgenerally that's in the public domain, news

(36:06):
worthy, public public information, informationthat the government voluntarily releases that you're free
to discuss and comment on, evenif it's about people. Especially if it's
about people. You don't get toyank that back out and say as a
regulator, no, you can't discussthis anymore. So that's one piece.
So that's not property at all.That belongs to everybody. So that's one

(36:28):
piece. The second piece is informationthat the government could regulate, but it
needs a reason, so in otherwords, because the first would get strict
to something called strict scrutiny, whichis the test that if you're the government's
on the government side of the vyou never want to see because you always

(36:50):
lose right, right, the shorthanding, shorthanding centuries of constitutional law. That
the middle pieces information that the governmentprobably could regulate if it does so in
a thoughtful and measured way. Andplus there may be policy reasons why you
want that information to circulate. Right, So, for example, financial information

(37:17):
about individuals, you would want incertain in certain circumstances to have people be
able to access that information, forexample, for looking at politically exposed persons,
or for money laundering purposes. Youwant to be able to exchange that
kind of information those kinds of things, or for credit. Right, you

(37:38):
could regulate that information. You justhave to do so in a reasonable way.
That's fine. And there's lots ofstuff, like there's lots of stuff
that falls in that bucket where thatif it's done in a measured way that
the potential the government can regulate thepotential for misuse because they're going to look

(37:58):
at the risk from that and decide, okay, no, you can't use
it for this. And then there'sa third tranch that's really easy to regulate,
where the where the harm is soobvious and intuitive from the use of
that information that it follows the waynight follows day. So for example,

(38:19):
unauthorized disrepute distribution of passwords or accountnumbers or full SSNs, et cetera,
like that type of stuff that hasa you know, yes, there are
no ideas there. Really it's akey of some kind. As an example,
I see, it's it's more functionalthan anything else. Right, That
kind of stuff is easy, anda commercial publisher of that type of information

(38:44):
is probably not serving a wildly usefulpurpose. Right, They're probably bad guys.
They're probably bad guys. But thereare there are people who maybe who
are you know, who may begiving a who may be selling geo geo
location data or other kinds of datathat have a lot of really valuable purposes

(39:05):
that you would want to allow tocontinue. And that is true both on
the private sector side. It's alsotrue on the government side, where the
availability of this kind of information isa legitimate concern, but it becomes a
proxy war for law enforcement priorities.So, in other words, rather than
saying, okay, you are usingcertain kinds of information to track down folks

(39:35):
who have overstayed their legal welcome inthe country, or if you're using it
to track down pregnant women those youknow, therefore we ought to ban all
law enforcement use of that information.Right, Right? Do you really want
that? Right? Right? Imean, do you really not want to

(39:57):
be able to get access to thecar tell operatives? Cell phone is that
the end result here, right,right, I mean those are that is
a completely legitimate discussion with the cliffs. These the cliffs they talk about,
right from a legal perspective, youcan go off the cliff this way,
you can got the cliff that way. You don't want to go off either
cliff. You want to stay somewherein the middle. Somewhere in the middle.

(40:22):
But that you know, it's avery charged environment around both of these
things, because there is no doubtthere are people, like there are people
doing bad things with data. There'sno doubt about that, right, I
mean, there are people selling thisto people that they shouldn't sell it to
I mean, our members don't.Our members don't do this that kind of
thing. That's why they I meanthey take know your customer type steps,

(40:45):
right, right, that's different.That's kind of part of the responsible action.
Yeah, but we view it asagain, if you view it through
a publishing lens, you it makesto me, it makes more sense.
It's in a sense easier to regulatebecause you're focusing on specific risks and then

(41:06):
you drill down on those risks andyou don't have to worry so much about
the First Amendment because you're going tobe narrower and more focused by definition.
Yeah, if you think of itas property that's being used willy nilly,
it's almost impossible. You're guaranteed yourchance of unintended consequences is it's almost guaranteed.
Yeah. Wow, that's excellent.Well, we've got time for one

(41:28):
more thread. I think the segment'sending here in a second, but I'm
going to tease this and we'll pickit up in the podcast bonus segment.
Okay, this whole concept of alternativedata, which now is everywhere and data
brokers are buying it and selling itall over the place. And I've done
some pretty good research on this,and I realized that there are several companies
I've talked to I know that arecapturing pretty much A lot of it is

(41:51):
geared around what they call exhaust datafrom credit card companies, but it's kind
of a misleading term because they willcollect how much you spent at this restaurant,
how much you spent at the gasstation. Like, all the detailed
transactions are actually being bought and soldto investment bankers, to other institutional investors,
and so what happens is did someof these folks have enough data now

(42:15):
to get a baseline and understand consumerbehavior. And they can't get cash,
but they can get credit cards.I don't think they can get debit cards,
but maybe some of them can.And the point being, there are
now organizations that know from the rawtransactional data at scale which companies are going
to meet or beat market estimates onWall Street. Well, if you know

(42:37):
that, then how the heck doyou lose when you gamble? Because the
analogy I give is, let's saythere's four of us playing cards at my
house, and the rule is Ican see all your cards, but none
of you can see my cards.Well, if I lose, I'm an
idiot. I don't know the rulesof gambling, apparently because I have such
an advantage now because I know whatall of you have and I know this

(42:59):
is happening. Argument. I'll letyou think about this as we good to
the break I thought was from apolicy perspective, Let's say that if a
company of credit card company is goingto sell this exhaust data to one of
these brokers, they must also publishanonymized data to some consumer facing data lake
that then any investor or any interestedparty could log into and kind of browse

(43:22):
around and see and better understand,because that kind of information is very valuable
if I run some kind of retailoperation and I can see at some scale
which products are selling, which productsaren't selling, what's happening in this region,
what's happening in that region. Thathelps me do a better job of
buying things that I know my customerswill probably want. Without revealing too much

(43:45):
about where the data came from,but at least it's useful, and what
I'm really angling at is a wayof leveling the playing field between these insiders.
And I even came up with aterm for what I call it,
outsider trading. So insider trading.Everyone understands that outsider trading is where from
the outside I've been able to gatherso much information that I know exactly what's

(44:05):
going to happen next. But we'llpick that up after the break in one
second. Be right back, allright, folks, were back here with
Chris Moore from SIIA. I justthrew a curveball question that him there.
I'm curious to know what are yourthoughts about all this alternative data stuff and
policies we can define that are reasonable. They will level of playing field.

(44:28):
So I so a couple of thingsthere. I mean, one is I
hear that word in a bunch ofdifferent that phrase rather in a bunch of
different contexts, and I think we'retalking about so I'm sure we're talking about
the same thing. What you're talkingabout is kind of information that's available that's

(44:51):
beyond that's outside of sort of thefour corners of required reporting, right.
That could be everything from as yousay, credit card exhaust credit company exhausts,
credit card exhaust data, to socialmedia trend data. And so when
this data is I think the waythe rules are now, this information is

(45:21):
scrubbed so that there's no material nonpublic information in it. So there was
a company called app Annie that gotinto a bit of hurt with the FTC,
with the SEC because it wasn't takingthat stuff out and lo and behold.
Like you can imagine, there areresults, right, were pretty good,

(45:44):
But that's not that's not what thebusiness is. What it is is
finding information from these different sources.You know, each of these firms has
their own way of balancing and analyzingit and then providing in as you said,

(46:05):
recommendations, trading recommendations to or datafrom which trades could be made to
right to uh, you know,to different types of investors. There is
at the moment. I mean,I think I would hesitate to mandate competition

(46:28):
in that space. I would beintuitively, I mean intuitively, I think
you know, the business is thatbusiness is I think growing. So in
other words, the initially it startedokay, we have we're getting this information

(46:50):
were as you said. I mean, a lot of this comes out of
what you said. Like the firstone was like all right, hey we
can do this. Let's do thisand sell it to hedge funds, right,
and that is like, you know, who else would be interested in
knowing all this stuff is, forexample, a retailer. Nothing wrong with
that either, because all of thisinformation is de identified, so it's really

(47:13):
just like there's no there's it's justokay, there were this many sales and
this is the these are the trendlines, you know, make your bets
right, right. And I thinkas the you know, as the business
evolves, my expectation is that it'sgoing to be that type of know how.
It's going to become more and morewidely available because there's no reason,

(47:37):
there's no reason for it not tobe interesting. That's a good point.
H Yeah, Well we're talking aboutthe information life cycle right and where it
is and where you can access itand what you can do with it.
And open data is a big thingthese days. I think that's wonderful news.
There's a lot of information coming out. I was actually an advocate way

(47:57):
back in two thousand and five whenI worked for the Data Warehousing Institute.
I got at a big soapbox andsaid, we need transparency in federal spending.
And I did this whole media campaign. I had Bacon's media source at
my access and so I emailed fortythousand reporters all about the need for transparency
and federal spending. Everyone told meI was crazy, And then something happened.

(48:20):
There was actually a guy at theHeritage Foundation, Mark Tapscott, who
picked up on it, and hewas like, why did you send me
this? And I was like,well, I went off my whole tangent.
He goes, I've been focused onthis for twenty years, Like where
have you been all my life?And he then he used my article and
really he was the imprimature of theData Warehousing Institute that he used to go
testify before Congress and say we needthese citizen auditors to help us find waste

(48:40):
and lo and behold, the Housepassed the bill. The Senate passed the
bill, co sponsored by a guynamed Barack Obama who was a Senator from
Illinois. And on September twenty sixth, two thousand and six, then President
George W. Bush signed the FederalFunding Accountability and Transparency Law. And I
almost had a heart attack on Myfriend called me and told me, I
was like, are you kidding me? No, I'm not going to say

(49:01):
I got any credit for or anything, which is fine, you know,
I don't mind, but I wasjust absolutely shocked. And now when you
look at it, you think mytheory back then was that, look,
everybody knows something. So even thoughsome accountant may not know that this line
on looks funky, there's someone whodoes. There's a carpenter, someone who
knows this kind of nail should onlycost ten cents, it should not cost

(49:22):
a dollar and ten cents, orwhatever the case may be. And these
days we have the capacity, likethe Amazon Mechanical Turk for example, to
leverage crowdsourcing at scale, and soif people register with the system, you
know, I'm a citizen. Youknow, everything I do in the system,
when I flag something, if I'mright or wrong, you can see

(49:43):
how you could dynamically score these thingsand then really wind up with a whole
cadre of experts greening or information systemsfor the government to help everyone know where
the money goes. Because a lotof times people worry about fraud. That
is true. Bad things do happen, are just mistakes, They're just things
that people overlook too. So beingable to find all that stuff, I

(50:05):
think it's very very interesting. Butfinal thoughts from you, Chris Moore of
s I I a here in theshow. Uh, you know, like
I said, I'm mostly optimistic aboutthese technologies. Yeah, you know,
do they need guardrails? Yeah,they do. They really do. Because
if if what you're trying to dois mimic human intelligence, you know,

(50:29):
human intelligence, we put guardrails aroundthat as well. There's things for not
allowed to do and we have goodreasons for doing that. Right, pretty
funny you go to put it.But the so you know, there's going
to be a need for that asthings develop, and what the type of
thing that you are discussing is appliedto the government is tremendously useful. I

(50:50):
think there are a number of privatebusinesses and even members who are selling all
kinds of solutions that do the samething. Right, they go into your
they go into your into your spreadheand they'll say, oh that's weird.
Yeah, you know, for ayear over here and employee changes one year
to the other. Maybe they don'tgo back five years. Maybe they're about
one year, right, you know, and chack to see how things have

(51:10):
changed, and suddenly that looks hotand it gets flat and that's how thanks
to tech fraud too. Yeah right, well, hey, this has been
this has been absolutely fantastic. Chrismore m HR. Look this organization up
online. SIIA doing wonderful work.You've made me more optimistic too about the
future. So it's very good toknow. And they're doing great work in
policy. Look them up online.Folks, you've been listening to Inside Analysis,

(51:35):
Go ahead and say it. We'reall ears CASEAA. Redlands Ranch Market
is a unique full service international orgrocery store that specializes in authentic food items
from Mexico, India, and frommany Mediterranean and Asian countries, including popular
items from the US. They offerfresh baked items from their in house bakery,
housemade tortillas from their Tortilla area,a delicious array of prepared Mexican foods,

(52:01):
a terrific fresh food and juice bar, and a large selection of meats,
seafoods and deli sandwiches, salads andhellal meats. Their produce department is
stocked full with fresh, local andhard to find international fruits and vegetables that
you cannot find anywhere else. Don'tforget to step into the massive beer cave
and experience the largest selection of domestic, artisan and imported beers in the ie.

(52:23):
They can also cater your next eventwith one of the delicious takeout catering
trades of food. Visit them atRedlands ranch Market dot com. That's Redlands
ranch Market dot com. Redlands RanchMarket a unique and fun shopping destination.
What is the Del Walmsley Radio Show? Welcome to the Del Walmsley Radio Show
where the high bends and the helpbegins. You need to stop being dependent

(52:46):
on a paycheck. All these selfhelp motivational people, they wind you up
like a little clock clill. Whois the show about. I'm your host,
Del Walmsley and all we're working onyour financial freedom. Learn the secrets
of building wealth from Dell Walmsley weekdaysfrom eleven am to noon right here on

(53:07):
KCAA to Hebot Club's original pure powerto rcosuper t comes from the only tree
in the world that fungus does notgrow on. As a result, it
naturally has anti fungal, anti infection, anti viral, antibacterial, anti inflammation,
and anti parasite properties. So thetea is great for healthy people because

(53:28):
it helps build the immune system,and it can be truly miraculous for someone
fighting a potentially life threatening disease dueto an infection, diabetes, or cancer.
The tea is also organic and naturallycaffeine free. A one pound package
of tea is forty nine to ninetyfive, which includes shipping. To order,
please visit to Hebota club dot com. T hebow is spelled tea like
tom, a, h ee blike boyo. Then continue with the word

(53:52):
t and then the word club.The complete website is to Hebot club dot
com or call us at eight oneeight six one zero eight zero eight eight
Monday through Saturday nine am to fivepm California time. That's eight one eight
six one zero eight zero eight eightto Hebot club dot Com. There's never
been a better time for men tobe whoever they want to be. Yet

(54:14):
it's never been less clear who menreally are. Guys Guy Radio, starring
author and Guy's Guy Robert Manny,is coming to Caseyaa every Wednesday at eight
pm. Whether it's relationships, sex, wellness for spirituality, join Robert as
he interviews the experts and shares hisinsights on building a world where men and

(54:34):
women can be at their best.Guys Guy Radio, Better Men, better
world. With sixty years of fascinatingfacts, this is the man from Yesterday
and back in time. We goto this time in nineteen sixty seven.
Frank Sinatra is not happy. Frankwants his girlfriend Mia Farrow to co star

(54:57):
in his new movie The Detective,but Pharaoh is doing Rosemary's Baby. That's
life. That's what all the peoplesay. You're riding high in April,
shut down in May. Back intime to this time in two thousand and
nine, Mariah Carey, Stevie Wonderand Usher sing emotional farewells to Michael Jackson

(55:22):
and described by his tearful eleven yearold daughter Paris as the best father you
could ever imagine. Ever since Iwas born, Daddy has been the best
father you can ever imagine. AndI just wanted to say I love him
so much and from this time innineteen ninety one. The Montell Williams Show

(55:43):
debuts on TV this week. Theshow will run thirteen weeks before any national
TV decision is made. Montell Williamsas a former Navy intelligence officer, this
is what I'm telling you. Ican't believe we told you. Listen,
we called him and said someone hasa crush. We didn't do it for
him. You went out and boughtNo matter who this person is, you

(56:04):
bought this person a set of diamondstud areas with more at man from Yesterday
dot com. Palm Springs Dispensary remindseveryone during these challenging times they understand the
importance of mental health. Offering dailydeals on everything from top quality flowers to
edibles and discounts on CBD products.Palm Springs Dispensary's goal is to help you

(56:27):
win the battle within. As oneof the area's most beautiful well inness shops.
Located on Garnet Avenue, visit thePalm Springs Dispensary dot com. That's
the Palm Springs Dispensary dot com experiencewhere expertise meets influence. On the insurance
hour with Carl Sussman. Carl's involvementwith California lawmakers and insurance discussions offers unparalleled
insights. Tune in for an experttake on the state's insurance landscape. KCAA

(56:53):
Radio has openings for one hour talkshows. If you want to host a
radio show, now is the time. Make KCIA your flangship station. Our
rates are affordable and our services aresecond to none. We broadcast to a
population of five million people plus.We stream and podcast on all major online
audio and video systems. If you'vebeen thinking about broadcasting a weekly radio program

(57:15):
on real radio plus the Internet,contact our CEO at two eight one five
nine nine ninety eight hundred two eightone five nine nine ninety eight hundred.
You can skype your show from yourhome to our Redlands, California studio,
where our live producers and engineers areready to work with you personally. A
radio program on KCIA is the perfectwork from home advocation in these stressful times.

(57:39):
Just type KCA radio dot com intoyour browser to learn more about hosting
a show. On the best stationin the nation, or call our CEO
for details too. Eight one,five nine nine ninety eight hundred NBC News
Radio. I'm Chris Gragio. PresidentBiden is leveling some of his harshest criticism

(57:59):
yet the Supreme Court. At afundraiser in Los Angeles yesterday. The President
called the conservative majority court out ofstep with the rest of the country.
Biden also warned that a former PresidentTrump is elected, he would appoint more
justices flying flags upside down, areference to conservative Justice Samuel Alito, who's
been embroiled in controversy over political flagsflown at his home. The comments were

(58:20):
made at a star studied fundraiser featuringformer President Obama and various A list celebrities.
Biden's campaign says the event raised arecord breaking thirty million dollars from Democratic
donors. Police saying i llegal immigrantis under arrest for the rape and murder
of a Maryland woman last August.In a news conference yesterday, Harford County
Sheriff Jeffrey Gaylor said a twenty threeyear old Victor Antonio Martinez Hernandez was arrested

(58:42):
Friday night in Tulsa, Oklahoma,and his waiting extradition. Rachel Moran,
a thirty seven year old mother offive, had been reported missing on August
fifth of last year, and herbody was found the next day. Maryland
Governor Wes Moore is saying it'll likelycost close to two billion dollars to rebuild
Baltimore's Francis Scott key Bridge. Inan interview with CBS's Face the Night,
he said, this state needs federalmoney to get the project completed in a

(59:02):
timely manner, so we know theestimated cost could be between one point seven
and one point nine billion dollars.And the reason that we are urging of
federal cost share and a one hundredpcent cost share is just simply because we
have to move quickly. The federalcost share means the federal government will cover
the entire cost. The governor alsopraised cleanup crews for their hard work reopening
the shipping channel ahead of schedule,saying they were able to reopen it in

(59:25):
eleven weeks despite estimates it would takeeleven months. Millions are celebrating Father's Day
today, but its origins are notwell known. It wasn't until nineteen seventy
two that Father's Day was recognized inAmerica after being signed off by President Nixon.
So Norah Dodd, the daughter ofa Civil War veteran, is credited
with organizing the first ever one backin nineteen ten to honor her own father,
and wanted the day to be inJune. I'm Chris Karagio, NBC

(59:49):
News Radio, NBC News on CACAALowland, sponsored by Teamsters Local nineteen thirty
two, Protecting the Future of WorkingFamilies Teamster nineteen thirty two dot org.
You're listening to an encore presentation ofthis