Ethical AI: Doing No Harm with Ivneet Kaur, Chief Information and Technology Officer at Sterling

[00:00:00] [SPEAKER_01]: Oh my goodness, bad touching, harassment, sex, violence, fraud, threats, all things that could

[00:00:10] [SPEAKER_01]: have been avoided.

[00:00:13] [SPEAKER_01]: If you had Fama, stop hiring dangerous people.

[00:00:19] [SPEAKER_01]: Fama.io

[00:00:20] [SPEAKER_02]: Hey, this is William Tinkup and Ryan Leary and you are listening and hopefully watching

[00:00:36] [SPEAKER_02]: the You Should Know podcast.

[00:00:39] [SPEAKER_02]: Ryan, how are you doing today?

[00:00:40] [SPEAKER_01]: Oh, I am wonderful.

[00:00:43] [SPEAKER_01]: I've never had a bad day in my life.

[00:00:45] [SPEAKER_01]: It's beautiful outside and...

[00:00:47] [SPEAKER_01]: Not true.

[00:00:47] [SPEAKER_01]: How do you know it's not true?

[00:00:49] [SPEAKER_01]: You don't live in my head.

[00:00:51] [SPEAKER_01]: I have evidence.

[00:00:52] [SPEAKER_01]: Well, that's kind of not true.

[00:00:53] [SPEAKER_01]: You do live in my head because I hear you at night.

[00:00:55] [SPEAKER_01]: I hear you in the morning.

[00:00:57] [SPEAKER_01]: Oh, that's terrible.

[00:00:58] [SPEAKER_02]: I don't even hear me.

[00:00:59] [SPEAKER_02]: I get tired of my own voice.

[00:01:02] [SPEAKER_02]: You know, Ivneet, how are you doing today?

[00:01:07] [SPEAKER_00]: I am doing very well, William and Ryan.

[00:01:10] [SPEAKER_00]: Yes, I have definitely had a bad day, Ryan, so I will not...

[00:01:14] [SPEAKER_02]: Although you're in Miami, you're not having too many bad days.

[00:01:19] [SPEAKER_00]: I'm not.

[00:01:21] [SPEAKER_00]: I can stare at the blue ocean right there.

[00:01:24] [SPEAKER_02]: That's not bad at all.

[00:01:27] [SPEAKER_02]: So while you're inside.

[00:01:29] [SPEAKER_02]: Yeah.

[00:01:30] [SPEAKER_02]: While we do some introduction, of course, tell us a little bit about yourself.

[00:01:39] [SPEAKER_00]: So from a professional perspective, I right now serve as the chief information and technology

[00:01:45] [SPEAKER_00]: officer for Sterling.

[00:01:48] [SPEAKER_00]: Any of you who don't know Sterling, we are a global provider of background and identity

[00:01:53] [SPEAKER_00]: services.

[00:01:54] [SPEAKER_00]: We serve clients in about 214 countries and hopefully you are also using us for hiring

[00:02:01] [SPEAKER_00]: people.

[00:02:01] [SPEAKER_00]: And I have responsibility for all of their internal and external technology data and

[00:02:09] [SPEAKER_00]: security ecosystem.

[00:02:11] [SPEAKER_00]: And I'm also the founder and sponsor of Sterling's AI program.

[00:02:15] [SPEAKER_02]: So, Ryan, before we get into questions, we've got a topic that we want to...

[00:02:19] [SPEAKER_02]: We're going to go down a lot of different rabbit holes, but with your expertise and

[00:02:24] [SPEAKER_02]: breadth of knowledge, we want to keep it simple and do the state of AI.

[00:02:28] [SPEAKER_02]: I just can't get your take on a lot of things that we're seeing in the marketplace.

[00:02:33] [SPEAKER_02]: So with that being said, let's just start basically because you're like you said,

[00:02:39] [SPEAKER_02]: you're the founder of kind of...

[00:02:41] [SPEAKER_02]: I would assume what you're trying to do is build ethical AI.

[00:02:47] [SPEAKER_02]: So how do you all...

[00:02:49] [SPEAKER_02]: What's your approach on auditing right now AI so that it's doing what it's

[00:02:56] [SPEAKER_02]: supposed to be doing?

[00:02:57] [SPEAKER_02]: So what's your take on that right now?

[00:03:03] [SPEAKER_00]: Yeah, so I think of it as three different monikers that I would say ethical,

[00:03:08] [SPEAKER_00]: responsible, secure.

[00:03:12] [SPEAKER_00]: So when we started our journey, we knew that there is a lot of hype around

[00:03:17] [SPEAKER_00]: things, but we didn't want to just jump into it and just start using it

[00:03:22] [SPEAKER_00]: and making it widely available and all those kind of things.

[00:03:26] [SPEAKER_00]: So we took a step back and we said, we are going to take a thoughtful

[00:03:30] [SPEAKER_00]: and holistic approach which takes into account what is our strategy?

[00:03:37] [SPEAKER_00]: What is the governance model?

[00:03:39] [SPEAKER_00]: How are we going to handle the pace of change?

[00:03:42] [SPEAKER_00]: How are we going to handle the upskilling?

[00:03:44] [SPEAKER_00]: And more importantly, how are we going to separate the hype from reality?

[00:03:51] [SPEAKER_00]: And there was a lot of do's and do nots which come into play there.

[00:03:55] [SPEAKER_00]: And part of that is also auditing and making sure that you're not

[00:04:00] [SPEAKER_00]: providing incorrect results and answers.

[00:04:04] [SPEAKER_00]: And that's where I think the thing which comes into play really is having

[00:04:07] [SPEAKER_00]: a very strong governance model, which is not just built on just

[00:04:13] [SPEAKER_00]: having engineers in it.

[00:04:14] [SPEAKER_00]: In this governance models, we have our HR partners, we have marketing,

[00:04:19] [SPEAKER_00]: we have privacy, we have compliance, we have security, we have product

[00:04:23] [SPEAKER_00]: and engineering.

[00:04:24] [SPEAKER_00]: So what we did was we set up this cross functional group who is going

[00:04:28] [SPEAKER_00]: to be able to be our steering committee and our governance group

[00:04:33] [SPEAKER_00]: to guide us as we bring use cases.

[00:04:36] [SPEAKER_00]: And we evaluate every single use case through multiple dimensions

[00:04:42] [SPEAKER_00]: and security, privacy and compliance are a huge part of it

[00:04:46] [SPEAKER_00]: to make sure is it something we should really be doing?

[00:04:49] [SPEAKER_00]: Should not be doing or can we do a portion of it?

[00:04:52] [SPEAKER_00]: So all of those things we take into consideration, even before we

[00:04:56] [SPEAKER_00]: start developing that use case.

[00:04:59] [SPEAKER_00]: Doesn't mean we are stopping experimentation, doesn't mean we

[00:05:01] [SPEAKER_00]: are not giving people sandboxes to do things.

[00:05:05] [SPEAKER_00]: But when it comes to real use case evaluation, we are being

[00:05:08] [SPEAKER_00]: very careful and thoughtful, especially because our business

[00:05:12] [SPEAKER_00]: is also heavily regulated.

[00:05:14] [SPEAKER_00]: People depend on us to hire their most important asset, people.

[00:05:19] [SPEAKER_00]: So for us, it's even more critical to do that and

[00:05:24] [SPEAKER_00]: you know, jumping in and just doing things without thinking

[00:05:28] [SPEAKER_00]: through all of those things is not always.

[00:05:30] [SPEAKER_01]: Yeah, how do we?

[00:05:32] [SPEAKER_01]: So we'll get into a number of these these scenarios as we

[00:05:37] [SPEAKER_01]: continue on. But how does one I want to take a break real

[00:05:43] [SPEAKER_01]: quick just to let you know about a new show.

[00:05:46] [SPEAKER_01]: We've just added to the network up next at work.

[00:05:50] [SPEAKER_01]: Hosted by Jean and Kate A'Keele of the Devon Group.

[00:05:54] [SPEAKER_01]: Fantastic show.

[00:05:56] [SPEAKER_01]: If you're looking for something that pushes the norm, pushes

[00:05:58] [SPEAKER_01]: the boundaries, has some really spirited conversations.

[00:06:02] [SPEAKER_01]: Google up next at work.

[00:06:05] [SPEAKER_01]: Jean and Kate A'Keele from the Devon Group.

[00:06:10] [SPEAKER_01]: Delineate between hype and reality.

[00:06:14] [SPEAKER_01]: I know that's a big question.

[00:06:16] [SPEAKER_01]: I get it.

[00:06:17] [SPEAKER_01]: She's going to take.

[00:06:19] [SPEAKER_01]: I know she's going to take on this.

[00:06:20] [SPEAKER_01]: But it was one of the first things you mentioned.

[00:06:24] [SPEAKER_01]: And and you know, it's kind of like, do I wait for this?

[00:06:26] [SPEAKER_01]: Let's just get into how do we delineate between hype and

[00:06:29] [SPEAKER_01]: reality because there's pressure from clients that want

[00:06:33] [SPEAKER_01]: you to be more innovative.

[00:06:36] [SPEAKER_02]: There's on the edge bleeding.

[00:06:38] [SPEAKER_01]: Right. Yeah.

[00:06:39] [SPEAKER_01]: On the bleeding edge is pressure from internal people

[00:06:41] [SPEAKER_01]: that said, why can't we go and do this?

[00:06:43] [SPEAKER_01]: How do we go faster?

[00:06:45] [SPEAKER_01]: How do we go further?

[00:06:46] [SPEAKER_01]: But how do we how do we hone in on that?

[00:06:49] [SPEAKER_01]: And really, how do you look at that?

[00:06:54] [SPEAKER_00]: So and this is going to sound very, very basic.

[00:06:58] [SPEAKER_00]: But it's these things are usually to do them.

[00:07:03] [SPEAKER_00]: They are simple, but they're always hard.

[00:07:06] [SPEAKER_00]: So the way I think about the North Star is the way

[00:07:09] [SPEAKER_00]: we bring back to people to people is OK, what is

[00:07:12] [SPEAKER_00]: the problem we're trying to solve for our customers?

[00:07:16] [SPEAKER_00]: Whether it's our internal employees or our external

[00:07:18] [SPEAKER_00]: customers, what is the problem?

[00:07:20] [SPEAKER_00]: Because I remember going through that even at

[00:07:23] [SPEAKER_00]: Sterling where the use cases were coming out of

[00:07:26] [SPEAKER_00]: woodwork, even though half of them did not need

[00:07:31] [SPEAKER_00]: AI to solve them.

[00:07:33] [SPEAKER_00]: But it just became came to people's mind.

[00:07:35] [SPEAKER_00]: Oh, this is this new AI Gen AI capabilities

[00:07:37] [SPEAKER_00]: available. Can I just use it to solve this

[00:07:40] [SPEAKER_00]: problem? The answer is no, not really.

[00:07:42] [SPEAKER_00]: There are cheaper, better ways to do it.

[00:07:45] [SPEAKER_00]: So but coming back to your question on hype

[00:07:48] [SPEAKER_00]: versus reality, first is your North Star is what

[00:07:51] [SPEAKER_00]: is the real business problem you're trying to solve?

[00:07:53] [SPEAKER_00]: Whether it's related to experience, efficiency,

[00:07:57] [SPEAKER_00]: optimization, productivity, whatever that is.

[00:07:59] [SPEAKER_00]: You have to have a very good sense of that.

[00:08:01] [SPEAKER_00]: Second, you have to have a really good evaluation

[00:08:04] [SPEAKER_00]: process and a group of smart people who are

[00:08:07] [SPEAKER_00]: able to look at these capabilities.

[00:08:09] [SPEAKER_00]: Oh, yeah, because there's a lot of AI washing

[00:08:11] [SPEAKER_00]: going on. I'm sure you guys have heard

[00:08:13] [SPEAKER_00]: on greenwashing. There is AI washing going

[00:08:16] [SPEAKER_00]: on where every single vendor has now added

[00:08:19] [SPEAKER_00]: AI as a buzzword to the offerings they do,

[00:08:22] [SPEAKER_00]: whether they use AI or not, or use it

[00:08:24] [SPEAKER_00]: in a very limited capacity.

[00:08:26] [SPEAKER_00]: So having the right talent in house

[00:08:29] [SPEAKER_00]: is very important also where somebody can dig

[00:08:32] [SPEAKER_00]: deep to test the

[00:08:35] [SPEAKER_00]: capability. So that testing and learning capability

[00:08:38] [SPEAKER_00]: from a technical perspective and from a privacy

[00:08:41] [SPEAKER_00]: and compliance perspective is important.

[00:08:44] [SPEAKER_00]: And that's sort of what we did. We put

[00:08:47] [SPEAKER_00]: a cross functional group of people together

[00:08:50] [SPEAKER_00]: to do that before we

[00:08:53] [SPEAKER_00]: before we go much further into it.

[00:08:56] [SPEAKER_00]: So those two things I would definitely

[00:08:59] [SPEAKER_00]: suggest that which happened and then there's

[00:09:01] [SPEAKER_00]: a continuous education.

[00:09:02] [SPEAKER_00]: So this is not a one and done thing.

[00:09:04] [SPEAKER_00]: The space is evolving incredibly fast.

[00:09:08] [SPEAKER_00]: It has, I would say the honeymoon period has

[00:09:10] [SPEAKER_00]: gotten over a little bit in 24 and people

[00:09:12] [SPEAKER_00]: are realizing that there is a lot more

[00:09:15] [SPEAKER_00]: to it than just launching some

[00:09:17] [SPEAKER_00]: chatbots if you will. And it requires

[00:09:20] [SPEAKER_00]: a more, you know, much bigger holistic approach.

[00:09:23] [SPEAKER_00]: But it's still a very cool innovative technology

[00:09:26] [SPEAKER_00]: and I am very I would say bullish about

[00:09:29] [SPEAKER_00]: there is a lot of potential to do a lot

[00:09:31] [SPEAKER_00]: of good in for the organization

[00:09:34] [SPEAKER_00]: and for people. So keep your not start

[00:09:36] [SPEAKER_00]: in mind what business problems you're solving.

[00:09:38] [SPEAKER_00]: Have the right set of skill sets in house

[00:09:41] [SPEAKER_00]: who can help you separate this hype from

[00:09:43] [SPEAKER_00]: reality and truly test.

[00:09:45] [SPEAKER_02]: Your customers probably have given you

[00:09:47] [SPEAKER_02]: this feedback as well because some things

[00:09:49] [SPEAKER_02]: that we hear in terms of centralizing

[00:09:52] [SPEAKER_02]: a I versus decentralizing

[00:09:55] [SPEAKER_02]: a I so

[00:09:56] [SPEAKER_02]: if some companies will have marketing will be

[00:09:59] [SPEAKER_02]: using gen AI to do X Y and Z

[00:10:01] [SPEAKER_02]: finances using it to do something else

[00:10:05] [SPEAKER_02]: and they wake up

[00:10:07] [SPEAKER_02]: and it's like AI is throughout the organization.

[00:10:09] [SPEAKER_02]: It's just not no one's got one finger

[00:10:12] [SPEAKER_02]: on the pulse of what's going on with AI,

[00:10:14] [SPEAKER_02]: which I think to your point creates kind

[00:10:16] [SPEAKER_02]: of security and privacy issues that could happen.

[00:10:20] [SPEAKER_02]: And it seems like you all have centralized it

[00:10:23] [SPEAKER_02]: to where if someone in marketing is going

[00:10:25] [SPEAKER_02]: to use AI, then they've got to go

[00:10:28] [SPEAKER_02]: through an approval process or some type of process.

[00:10:31] [SPEAKER_02]: Did I get that right?

[00:10:34] [SPEAKER_00]: Yes, yes, yes.

[00:10:36] [SPEAKER_00]: So I think of it

[00:10:38] [SPEAKER_00]: as you have to centralize to decentralize.

[00:10:41] [SPEAKER_00]: So when you're starting on these things

[00:10:43] [SPEAKER_00]: from a journey perspective, you have to

[00:10:45] [SPEAKER_00]: make sure you're giving people the right

[00:10:47] [SPEAKER_00]: guardrails, the right set of tools,

[00:10:49] [SPEAKER_00]: the right set of policies, the right set

[00:10:51] [SPEAKER_00]: of processes, a combination of all these

[00:10:54] [SPEAKER_00]: things and not to mention the organizational

[00:10:56] [SPEAKER_00]: change management.

[00:10:58] [SPEAKER_00]: But then you're also giving them

[00:10:59] [SPEAKER_00]: enough freedom within the guardrails

[00:11:02] [SPEAKER_00]: to experiment and launch

[00:11:05] [SPEAKER_00]: their own capabilities.

[00:11:07] [SPEAKER_00]: So it's I think of it a little bit more

[00:11:09] [SPEAKER_00]: of a hybrid model, if you will,

[00:11:12] [SPEAKER_00]: not fully centralized and not fully decentralized,

[00:11:15] [SPEAKER_00]: especially when you're starting the

[00:11:16] [SPEAKER_00]: centralization is very critical

[00:11:18] [SPEAKER_00]: to do to set the right guardrails.

[00:11:20] [SPEAKER_02]: Especially again, letting people innovate,

[00:11:23] [SPEAKER_02]: you don't want to you don't want to squish

[00:11:24] [SPEAKER_02]: that out of any department or any area

[00:11:26] [SPEAKER_02]: in the company.

[00:11:27] [SPEAKER_02]: Everyone should be able to innovate.

[00:11:29] [SPEAKER_02]: But at the same time, if the innovation

[00:11:31] [SPEAKER_02]: opens you up to a security breach

[00:11:34] [SPEAKER_02]: or privacy breach, the innovations

[00:11:36] [SPEAKER_02]: not worth it.

[00:11:38] [SPEAKER_02]: So I mean, there is a balancing that

[00:11:40] [SPEAKER_02]: you're delicate balancing that

[00:11:41] [SPEAKER_02]: professionals like you are trying to

[00:11:43] [SPEAKER_02]: figure out like, OK,

[00:11:45] [SPEAKER_02]: we need innovation at the same time.

[00:11:47] [SPEAKER_02]: Data's got to be secure, it's got to be private.

[00:11:50] [SPEAKER_02]: I mean, yeah.

[00:11:51] [SPEAKER_01]: One of the questions I have

[00:11:53] [SPEAKER_01]: is we hear a lot about ethical

[00:11:55] [SPEAKER_01]: AI and ethical everything.

[00:11:59] [SPEAKER_01]: But from the horse's mouth, someone

[00:12:01] [SPEAKER_01]: who's doing this at a large company

[00:12:04] [SPEAKER_01]: and doing and managing overseeing a lot

[00:12:07] [SPEAKER_01]: of innovation.

[00:12:08] [SPEAKER_01]: What does ethical AI mean to you?

[00:12:12] [SPEAKER_02]: Yeah, it's a great question.

[00:12:14] [SPEAKER_02]: Because is ethical subjective?

[00:12:19] [SPEAKER_00]: It could be. But.

[00:12:22] [SPEAKER_00]: Ethical to me is do no harm.

[00:12:24] [SPEAKER_00]: OK. That's do no harm.

[00:12:27] [SPEAKER_00]: That's the definition I go with.

[00:12:29] [SPEAKER_00]: Well, you know, overall in general,

[00:12:31] [SPEAKER_00]: not just in context of using AI

[00:12:34] [SPEAKER_00]: or any technology or building any

[00:12:36] [SPEAKER_00]: product. So we don't want to

[00:12:38] [SPEAKER_00]: be biased, let's say towards

[00:12:41] [SPEAKER_00]: candidates, whether we are hiring them

[00:12:42] [SPEAKER_00]: internally or helping our clients hire

[00:12:44] [SPEAKER_00]: them externally.

[00:12:46] [SPEAKER_00]: We don't want to give our clients

[00:12:48] [SPEAKER_00]: something which can make them

[00:12:50] [SPEAKER_00]: biased towards a certain set

[00:12:52] [SPEAKER_00]: of candidate pool they're trying

[00:12:54] [SPEAKER_00]: to hire. So that's the lens

[00:12:56] [SPEAKER_00]: to use that how do we remove

[00:12:59] [SPEAKER_00]: bias and how do we not harm

[00:13:01] [SPEAKER_00]: our clients or our candidates

[00:13:04] [SPEAKER_00]: as we are developing any products

[00:13:06] [SPEAKER_00]: or solutions in AI

[00:13:08] [SPEAKER_00]: just exasperates that

[00:13:10] [SPEAKER_00]: more because of, you know,

[00:13:12] [SPEAKER_00]: the capabilities that it brings to

[00:13:15] [SPEAKER_00]: the table.

[00:13:16] [SPEAKER_00]: But yeah, that's how I think we'll

[00:13:18] [SPEAKER_02]: steal that if you don't mind.

[00:13:19] [SPEAKER_02]: We're going to steal that and use

[00:13:21] [SPEAKER_02]: that for ourselves.

[00:13:23] [SPEAKER_02]: All right. So Ryan and I do a new

[00:13:25] [SPEAKER_02]: show on Sundays and we cover

[00:13:27] [SPEAKER_02]: the market

[00:13:28] [SPEAKER_02]: a couple of weeks ago in Dresden

[00:13:31] [SPEAKER_02]: Horowitz came out and

[00:13:32] [SPEAKER_02]: said they were going to support Trump.

[00:13:35] [SPEAKER_02]: But one of the reason that

[00:13:37] [SPEAKER_02]: Mark said he's going to support

[00:13:39] [SPEAKER_02]: Trump was

[00:13:41] [SPEAKER_02]: he doesn't want government interference

[00:13:43] [SPEAKER_02]: with AI, which I found fascinating.

[00:13:46] [SPEAKER_02]: There's just a whole thing that's

[00:13:47] [SPEAKER_02]: fascinating on some some level.

[00:13:49] [SPEAKER_02]: Right. So what's your take

[00:13:51] [SPEAKER_02]: on because y'all are heavily

[00:13:53] [SPEAKER_02]: regulated industry, right?

[00:13:56] [SPEAKER_02]: Could almost like finance or like

[00:13:58] [SPEAKER_02]: insurance, health care, etc.

[00:14:00] [SPEAKER_02]: Like, OK, got it.

[00:14:03] [SPEAKER_02]: What do you think about government

[00:14:05] [SPEAKER_02]: jumping into this particular

[00:14:07] [SPEAKER_02]: foray into this particular area

[00:14:11] [SPEAKER_02]: and setting regs and setting

[00:14:13] [SPEAKER_02]: different things that you can do,

[00:14:14] [SPEAKER_02]: can't do, etc.

[00:14:15] [SPEAKER_02]: Like is that a is that

[00:14:18] [SPEAKER_02]: a positive thing for your net net?

[00:14:19] [SPEAKER_02]: Or is that you'd rather

[00:14:21] [SPEAKER_02]: like Mr. Horowitz?

[00:14:24] [SPEAKER_00]: I think it's a positive thing.

[00:14:26] [SPEAKER_00]: I personally think government has

[00:14:28] [SPEAKER_00]: a role to play.

[00:14:30] [SPEAKER_00]: I do think government has a role to play.

[00:14:32] [SPEAKER_00]: And here is here is how I think about

[00:14:35] [SPEAKER_00]: it. And again, I'm going to

[00:14:37] [SPEAKER_00]: see it in broad terms.

[00:14:39] [SPEAKER_00]: The the the way

[00:14:41] [SPEAKER_00]: I think about it is that you're

[00:14:43] [SPEAKER_00]: getting two different lenses.

[00:14:45] [SPEAKER_00]: Government has a different lens

[00:14:46] [SPEAKER_00]: to this problem than Mark does.

[00:14:50] [SPEAKER_00]: Mark is trying to drive the innovation

[00:14:52] [SPEAKER_00]: really fast and solve very,

[00:14:54] [SPEAKER_00]: very big problems.

[00:14:56] [SPEAKER_00]: Yes, got it.

[00:14:57] [SPEAKER_00]: That that is critical to do.

[00:14:59] [SPEAKER_00]: And you know what what has happened

[00:15:01] [SPEAKER_00]: in Silicon Valley has changed all our

[00:15:03] [SPEAKER_00]: lives and words.

[00:15:04] [SPEAKER_00]: So big fan of being able to

[00:15:06] [SPEAKER_00]: move fast.

[00:15:08] [SPEAKER_00]: But I'm not I'm not a big believer

[00:15:10] [SPEAKER_00]: in moving fast and breaking so much

[00:15:12] [SPEAKER_00]: that we start harming people.

[00:15:14] [SPEAKER_00]: And that's where I start.

[00:15:15] [SPEAKER_00]: I see the government's role coming

[00:15:17] [SPEAKER_00]: in where you're looking at a holistic

[00:15:19] [SPEAKER_00]: ecosystem and you're trying to

[00:15:21] [SPEAKER_00]: prevent harm to

[00:15:23] [SPEAKER_00]: the most vulnerable people.

[00:15:25] [SPEAKER_00]: So look at the use of AI,

[00:15:26] [SPEAKER_00]: which can happen in people

[00:15:28] [SPEAKER_00]: getting scammed.

[00:15:30] [SPEAKER_00]: And the scams are becoming more

[00:15:31] [SPEAKER_00]: realistic and that's just

[00:15:32] [SPEAKER_00]: Nigerian Prince email

[00:15:34] [SPEAKER_00]: those things that we do have

[00:15:35] [SPEAKER_02]: regulation in Nigeria and Prince

[00:15:36] [SPEAKER_02]: email that we used to get.

[00:15:39] [SPEAKER_02]: I'm more sophisticated.

[00:15:41] [SPEAKER_00]: Now it's like

[00:15:42] [SPEAKER_00]: my mom will send me an email and I

[00:15:45] [SPEAKER_00]: will not be able to recognize that

[00:15:46] [SPEAKER_00]: it's a fake.

[00:15:48] [SPEAKER_00]: So those kind of things,

[00:15:50] [SPEAKER_00]: that's where I think the government

[00:15:51] [SPEAKER_00]: plays a role because it looks at

[00:15:53] [SPEAKER_00]: a much wider spectrum

[00:15:56] [SPEAKER_00]: of people

[00:15:58] [SPEAKER_00]: to make sure that we have

[00:16:00] [SPEAKER_00]: the regulations and protections

[00:16:01] [SPEAKER_00]: in place while the innovation

[00:16:03] [SPEAKER_00]: is happening. The problem is

[00:16:05] [SPEAKER_00]: government is always doing

[00:16:06] [SPEAKER_00]: catching up.

[00:16:07] [SPEAKER_00]: That's what frustrates the

[00:16:09] [SPEAKER_00]: innovators because government

[00:16:10] [SPEAKER_00]: is like five years behind

[00:16:12] [SPEAKER_00]: in catching up.

[00:16:14] [SPEAKER_00]: So to me, the problem

[00:16:17] [SPEAKER_00]: to solve is how do you bring

[00:16:18] [SPEAKER_00]: government along quicker?

[00:16:20] [SPEAKER_00]: How do you have that think

[00:16:22] [SPEAKER_00]: tank that group of people

[00:16:24] [SPEAKER_00]: who are

[00:16:26] [SPEAKER_00]: working together from the very

[00:16:28] [SPEAKER_00]: beginning when a new tech is

[00:16:30] [SPEAKER_00]: getting developed and what its

[00:16:31] [SPEAKER_00]: impact can be versus

[00:16:33] [SPEAKER_00]: trying to be at odds with each

[00:16:35] [SPEAKER_00]: other because I believe they

[00:16:36] [SPEAKER_00]: can be really great partners

[00:16:37] [SPEAKER_00]: together to really drive

[00:16:39] [SPEAKER_00]: this innovation in a responsible

[00:16:41] [SPEAKER_01]: way. Then we get into a way

[00:16:43] [SPEAKER_01]: too political for me, but

[00:16:44] [SPEAKER_01]: that's when we get into Congress

[00:16:45] [SPEAKER_01]: and the Senate and being

[00:16:48] [SPEAKER_01]: tech innovators.

[00:16:49] [SPEAKER_01]: Right. And I think

[00:16:51] [SPEAKER_01]: they question Zuckerberg and

[00:16:53] [SPEAKER_01]: some of these others. And I

[00:16:54] [SPEAKER_01]: just I listen to it.

[00:16:56] [SPEAKER_01]: Man, this is awkward.

[00:16:58] [SPEAKER_01]: Like listening to tell me about

[00:17:00] [SPEAKER_02]: this my book.

[00:17:01] [SPEAKER_01]: Yeah. Yeah. What is my blog?

[00:17:03] [SPEAKER_01]: Right. Yeah.

[00:17:04] [SPEAKER_01]: So.

[00:17:06] [SPEAKER_01]: So you have face space.

[00:17:09] [SPEAKER_01]: What do you do on the face?

[00:17:10] [SPEAKER_03]: Space. Yeah.

[00:17:11] [SPEAKER_01]: Yeah. So it's awkward.

[00:17:13] [SPEAKER_01]: Yeah.

[00:17:14] [SPEAKER_01]: But I need I like your

[00:17:17] [SPEAKER_01]: take on that. I think that's

[00:17:18] [SPEAKER_01]: I didn't expect it.

[00:17:20] [SPEAKER_02]: I did not expect that actually.

[00:17:21] [SPEAKER_02]: It's very easy to understand.

[00:17:23] [SPEAKER_02]: I have to admit I thought

[00:17:24] [SPEAKER_02]: you're going to go the other

[00:17:25] [SPEAKER_02]: direction.

[00:17:26] [SPEAKER_02]: Indeed.

[00:17:27] [SPEAKER_02]: I thought you were going to go

[00:17:28] [SPEAKER_02]: the extra key from out of our

[00:17:29] [SPEAKER_02]: business. Just keep them out.

[00:17:31] [SPEAKER_02]: Let us do our bit.

[00:17:32] [SPEAKER_02]: We'll manage it.

[00:17:33] [SPEAKER_01]: But it does make sense, right?

[00:17:34] [SPEAKER_01]: I mean, technology moves

[00:17:36] [SPEAKER_01]: incredibly fast, right?

[00:17:38] [SPEAKER_01]: And and these

[00:17:39] [SPEAKER_01]: these big companies who have

[00:17:41] [SPEAKER_01]: leaders that otherwise

[00:17:43] [SPEAKER_01]: we would be significantly behind

[00:17:45] [SPEAKER_01]: the rest of the world are

[00:17:47] [SPEAKER_01]: pushing at that

[00:17:48] [SPEAKER_01]: that type of speed.

[00:17:50] [SPEAKER_01]: They don't want to be held

[00:17:51] [SPEAKER_01]: back.

[00:17:52] [SPEAKER_01]: But then we're left to.

[00:17:55] [SPEAKER_01]: The feelings of that person and

[00:17:57] [SPEAKER_01]: what they want to do to

[00:17:58] [SPEAKER_01]: accomplish their goal.

[00:17:59] [SPEAKER_01]: So the oversight, how you explain

[00:18:01] [SPEAKER_01]: that I think makes complete

[00:18:02] [SPEAKER_01]: complete sense.

[00:18:04] [SPEAKER_01]: Question I have

[00:18:06] [SPEAKER_01]: is around the state

[00:18:08] [SPEAKER_01]: of AI.

[00:18:11] [SPEAKER_01]: If you had to and I know

[00:18:13] [SPEAKER_01]: this is maybe a really bad

[00:18:15] [SPEAKER_01]: question because you can't

[00:18:17] [SPEAKER_01]: sum it up in a couple of

[00:18:18] [SPEAKER_01]: sentences, although I think

[00:18:19] [SPEAKER_01]: you might be able to do this.

[00:18:20] [SPEAKER_01]: You feel I feel like you have

[00:18:22] [SPEAKER_01]: this ability.

[00:18:23] [SPEAKER_01]: How would you describe

[00:18:25] [SPEAKER_01]: or explain the state of AI to

[00:18:27] [SPEAKER_01]: somebody who's

[00:18:29] [SPEAKER_01]: just casually interested

[00:18:31] [SPEAKER_01]: in what's happening?

[00:18:37] [SPEAKER_01]: Oh, my goodness.

[00:18:39] [SPEAKER_01]: Bad touching, harassment,

[00:18:41] [SPEAKER_01]: sex, violence,

[00:18:43] [SPEAKER_01]: fraud, threats,

[00:18:45] [SPEAKER_01]: all things that could

[00:18:47] [SPEAKER_01]: have been avoided

[00:18:50] [SPEAKER_01]: if you had Fama

[00:18:52] [SPEAKER_01]: stop hiring dangerous

[00:18:55] [SPEAKER_01]: people.

[00:18:56] [SPEAKER_01]: Fama.io.

[00:18:59] [SPEAKER_02]: I while you're thinking, I'll

[00:19:01] [SPEAKER_02]: tell you what I read the

[00:19:03] [SPEAKER_02]: other day and Ryan, I've told

[00:19:04] [SPEAKER_02]: you this.

[00:19:06] [SPEAKER_02]: Guy who's an analyst and

[00:19:09] [SPEAKER_02]: he said we're in the first

[00:19:11] [SPEAKER_02]: inning of a 10,000

[00:19:13] [SPEAKER_02]: inning game.

[00:19:14] [SPEAKER_02]: So what what we

[00:19:16] [SPEAKER_02]: what we know of AI

[00:19:18] [SPEAKER_02]: again, the difference between

[00:19:19] [SPEAKER_02]: machine learning and LP and

[00:19:22] [SPEAKER_02]: large language models and all

[00:19:23] [SPEAKER_02]: these different things that we

[00:19:26] [SPEAKER_02]: kind of understand.

[00:19:29] [SPEAKER_02]: We have no in our lifetimes

[00:19:31] [SPEAKER_02]: will never know.

[00:19:32] [SPEAKER_02]: Well, we won't see a fraction of

[00:19:34] [SPEAKER_02]: what's to be developed.

[00:19:35] [SPEAKER_02]: It's the way that he kind of

[00:19:37] [SPEAKER_02]: the way he packaged up.

[00:19:40] [SPEAKER_00]: I agree.

[00:19:42] [SPEAKER_00]: I agree. We won't.

[00:19:43] [SPEAKER_00]: We won't. And that's what makes

[00:19:45] [SPEAKER_00]: it exciting to me.

[00:19:46] [SPEAKER_00]: So to me, if somebody is

[00:19:47] [SPEAKER_00]: casually browsing, I would say

[00:19:49] [SPEAKER_00]: yeah, um, learn

[00:19:51] [SPEAKER_00]: about it. It's exciting.

[00:19:53] [SPEAKER_00]: This is the time to learn about

[00:19:55] [SPEAKER_00]: it and get yourself familiar with

[00:19:56] [SPEAKER_00]: it. You are already using it.

[00:19:58] [SPEAKER_00]: So if I'm a casual person like

[00:19:59] [SPEAKER_00]: if I use Amazon, everybody uses

[00:20:01] [SPEAKER_00]: Amazon.

[00:20:02] [SPEAKER_00]: I am using AI at this time

[00:20:04] [SPEAKER_00]: because that's what they're

[00:20:05] [SPEAKER_00]: using to give me recommendations

[00:20:07] [SPEAKER_00]: and even their own

[00:20:09] [SPEAKER_00]: operation.

[00:20:10] [SPEAKER_00]: So learn about it.

[00:20:12] [SPEAKER_00]: Be curious about it.

[00:20:14] [SPEAKER_00]: Because and I think of it

[00:20:16] [SPEAKER_00]: at, you know, from the other

[00:20:18] [SPEAKER_00]: dimensions, if if you're a

[00:20:19] [SPEAKER_00]: parent, you want to think about

[00:20:20] [SPEAKER_00]: it. How can it impact your kids?

[00:20:23] [SPEAKER_00]: If you're a student, how is it

[00:20:24] [SPEAKER_00]: going to impact your future job?

[00:20:26] [SPEAKER_00]: If you are a professional like

[00:20:27] [SPEAKER_00]: me, you have no choice but to

[00:20:29] [SPEAKER_00]: continue to learn about

[00:20:32] [SPEAKER_00]: it. So

[00:20:33] [SPEAKER_00]: everybody has something to gain

[00:20:35] [SPEAKER_00]: by learning about it.

[00:20:37] [SPEAKER_00]: So how do we simplify that?

[00:20:39] [SPEAKER_00]: How do we simplify it for

[00:20:40] [SPEAKER_00]: different personas

[00:20:42] [SPEAKER_00]: with different age groups

[00:20:43] [SPEAKER_00]: and make it more targeted on

[00:20:45] [SPEAKER_00]: how it impacts them is what

[00:20:47] [SPEAKER_00]: I would love

[00:20:49] [SPEAKER_00]: to see more focus on.

[00:20:51] [SPEAKER_00]: Because there is the information

[00:20:52] [SPEAKER_00]: is being curious.

[00:20:54] [SPEAKER_00]: There is so much information

[00:20:55] [SPEAKER_02]: out there. Ryan uses

[00:20:58] [SPEAKER_02]: chat, GBT.

[00:21:00] [SPEAKER_02]: What's version of you on your own

[00:21:01] [SPEAKER_02]: for four? Yeah.

[00:21:03] [SPEAKER_02]: Yeah. And it's it's funny

[00:21:05] [SPEAKER_02]: because like every day he'll

[00:21:06] [SPEAKER_02]: bring some new prompt or a new

[00:21:07] [SPEAKER_02]: thing that he's learning.

[00:21:09] [SPEAKER_02]: And it reminds me very much of

[00:21:11] [SPEAKER_02]: the internet in like say

[00:21:13] [SPEAKER_02]: 1998

[00:21:15] [SPEAKER_02]: where it was just fun.

[00:21:17] [SPEAKER_02]: Like you get on the web and

[00:21:19] [SPEAKER_02]: you know the message boards and

[00:21:20] [SPEAKER_02]: like it was just a fun bit.

[00:21:22] [SPEAKER_02]: Like you didn't think about all

[00:21:23] [SPEAKER_02]: the downsides and all the

[00:21:24] [SPEAKER_02]: darkness and all the other stuff

[00:21:26] [SPEAKER_02]: that would come or there

[00:21:27] [SPEAKER_02]: probably even was there.

[00:21:29] [SPEAKER_02]: But you just had fun.

[00:21:30] [SPEAKER_02]: You're like, OK, let me see if I

[00:21:31] [SPEAKER_02]: can find a pet store in

[00:21:32] [SPEAKER_02]: Pittsburgh.

[00:21:33] [SPEAKER_02]: Like you just do this random

[00:21:35] [SPEAKER_02]: stuff just to see it

[00:21:37] [SPEAKER_02]: it was playful.

[00:21:39] [SPEAKER_02]: And again, nobody got hurt.

[00:21:40] [SPEAKER_02]: It was just it was fun.

[00:21:43] [SPEAKER_02]: I see the parallels.

[00:21:45] [SPEAKER_02]: I think it's going to move far

[00:21:47] [SPEAKER_02]: faster than the internet did.

[00:21:50] [SPEAKER_02]: So I think that Moore's law

[00:21:51] [SPEAKER_02]: applied to AI versus

[00:21:54] [SPEAKER_02]: the internet. I think that's going

[00:21:56] [SPEAKER_02]: to happen.

[00:21:59] [SPEAKER_00]: I do think more people who learn

[00:22:01] [SPEAKER_00]: about it that will

[00:22:03] [SPEAKER_00]: make

[00:22:05] [SPEAKER_00]: that will make innovation faster

[00:22:07] [SPEAKER_00]: but also safer because

[00:22:09] [SPEAKER_00]: more people who know about it

[00:22:10] [SPEAKER_00]: then you have more perspectives

[00:22:12] [SPEAKER_00]: coming in both positive and

[00:22:14] [SPEAKER_00]: negative. And that's what

[00:22:16] [SPEAKER_00]: all the innovators in Silicon

[00:22:18] [SPEAKER_00]: Valley and worldwide need to

[00:22:20] [SPEAKER_00]: hear that rather than

[00:22:21] [SPEAKER_00]: you know, very close group of

[00:22:23] [SPEAKER_00]: people

[00:22:25] [SPEAKER_00]: who are at the cutting edge of

[00:22:27] [SPEAKER_00]: it. But that's where the bias

[00:22:29] [SPEAKER_00]: also comes in because you're not

[00:22:30] [SPEAKER_00]: really hearing from everybody

[00:22:32] [SPEAKER_00]: you need to hear.

[00:22:32] [SPEAKER_02]: OK, so you've probably seen this

[00:22:34] [SPEAKER_02]: in different cycles, but

[00:22:35] [SPEAKER_02]: I've always been curious about

[00:22:37] [SPEAKER_02]: how standards get made

[00:22:39] [SPEAKER_02]: like, like, you know, like

[00:22:41] [SPEAKER_02]: simple stuff like an iPhone

[00:22:44] [SPEAKER_02]: plug versus a droid plug,

[00:22:46] [SPEAKER_02]: right? Like who made the decision

[00:22:47] [SPEAKER_02]: and why didn't they just make the

[00:22:49] [SPEAKER_02]: same damn plug?

[00:22:50] [SPEAKER_02]: My house would be so much smoother.

[00:22:53] [SPEAKER_02]: But you know, like it's like big

[00:22:54] [SPEAKER_02]: companies

[00:22:56] [SPEAKER_02]: they'll make decisions on standards

[00:22:58] [SPEAKER_02]: that that are their standards

[00:23:00] [SPEAKER_02]: and and do you see

[00:23:02] [SPEAKER_02]: that play out?

[00:23:04] [SPEAKER_02]: I mean, obviously the big ones with

[00:23:06] [SPEAKER_02]: Microsoft and Google and Facebook

[00:23:08] [SPEAKER_02]: and pick all your other ones

[00:23:10] [SPEAKER_02]: that you like.

[00:23:11] [SPEAKER_02]: Is it going to become kind of a

[00:23:14] [SPEAKER_02]: another game of them creating

[00:23:15] [SPEAKER_02]: standards and adhering

[00:23:18] [SPEAKER_02]: to

[00:23:20] [SPEAKER_02]: the standard of a company

[00:23:22] [SPEAKER_02]: versus general

[00:23:24] [SPEAKER_02]: standard of, you know, AI

[00:23:27] [SPEAKER_02]: is versus AI

[00:23:29] [SPEAKER_02]: with Microsoft is AI

[00:23:31] [SPEAKER_02]: with Google is AI with Amazon

[00:23:33] [SPEAKER_02]: is, etc.

[00:23:34] [SPEAKER_02]: So question is about standards.

[00:23:37] [SPEAKER_00]: I don't have a

[00:23:39] [SPEAKER_00]: Yeah, there will be some of that.

[00:23:41] [SPEAKER_00]: I will tell you and maybe I'm

[00:23:42] [SPEAKER_00]: glass half full

[00:23:44] [SPEAKER_00]: on this one. But I do believe

[00:23:47] [SPEAKER_00]: there will be a lot more

[00:23:48] [SPEAKER_00]: interoperability.

[00:23:49] [SPEAKER_02]: Well, you are an optimist.

[00:23:50] [SPEAKER_02]: Wow.

[00:23:51] [SPEAKER_02]: I like this.

[00:23:52] [SPEAKER_02]: Wow. All righty, we found one.

[00:23:54] [SPEAKER_02]: Right? We were looking.

[00:23:56] [SPEAKER_01]: I can see them all next week.

[00:23:57] [SPEAKER_01]: They're all around a table.

[00:23:59] [SPEAKER_01]: Like, let's get standards for

[00:24:01] [SPEAKER_01]: everybody.

[00:24:03] [SPEAKER_00]: Maybe not.

[00:24:04] [SPEAKER_00]: Maybe not.

[00:24:05] [SPEAKER_00]: They all want to push their

[00:24:07] [SPEAKER_00]: products. But

[00:24:09] [SPEAKER_00]: however, but I do believe they

[00:24:10] [SPEAKER_00]: also want interoperability

[00:24:12] [SPEAKER_00]: because they also know that one

[00:24:13] [SPEAKER_00]: of them is not going to

[00:24:16] [SPEAKER_00]: what encode when this race if

[00:24:17] [SPEAKER_00]: you OK. So I'll give you a

[00:24:18] [SPEAKER_00]: parallel example, which

[00:24:20] [SPEAKER_00]: hopefully they all learn from.

[00:24:21] [SPEAKER_00]: We have three

[00:24:24] [SPEAKER_00]: main cloud providers.

[00:24:25] [SPEAKER_00]: You have GCP, you have AWS,

[00:24:27] [SPEAKER_00]: you have Azure.

[00:24:29] [SPEAKER_00]: OK, it's not that easy

[00:24:30] [SPEAKER_00]: to move from one to another.

[00:24:33] [SPEAKER_00]: But more and more companies

[00:24:35] [SPEAKER_00]: want to do that from a

[00:24:37] [SPEAKER_00]: DR perspective, BCP perspective,

[00:24:40] [SPEAKER_00]: capabilities perspective.

[00:24:42] [SPEAKER_00]: So to me that increases

[00:24:44] [SPEAKER_00]: the usage of cloud and the cloud

[00:24:46] [SPEAKER_00]: adoption and all those things.

[00:24:47] [SPEAKER_00]: If you actually make that happen,

[00:24:49] [SPEAKER_00]: so it's a good business.

[00:24:51] [SPEAKER_00]: So it's not. Yeah, yeah, yeah.

[00:24:52] [SPEAKER_00]: Goodness of the heart, but it's

[00:24:54] [SPEAKER_00]: good business.

[00:24:55] [SPEAKER_00]: That's how I think about the AI

[00:24:57] [SPEAKER_00]: journey as well, that hopefully

[00:24:59] [SPEAKER_00]: learning from those parallels,

[00:25:01] [SPEAKER_00]: these companies are going to be

[00:25:02] [SPEAKER_00]: more open

[00:25:06] [SPEAKER_00]: to having at least a set of

[00:25:08] [SPEAKER_00]: base standards where they are

[00:25:09] [SPEAKER_00]: interoperable and they're able

[00:25:11] [SPEAKER_00]: to use each other stuff or allow

[00:25:13] [SPEAKER_00]: clients like us to build things

[00:25:16] [SPEAKER_00]: on top of two providers

[00:25:17] [SPEAKER_00]: or have, you know, a stronger

[00:25:20] [SPEAKER_00]: DR plan.

[00:25:21] [SPEAKER_00]: So that's at least the parallel

[00:25:23] [SPEAKER_00]: I draw in my head when I think

[00:25:24] [SPEAKER_00]: about it.

[00:25:25] [SPEAKER_00]: Of course, they all get very,

[00:25:26] [SPEAKER_00]: very big bucks to make those

[00:25:28] [SPEAKER_00]: decisions.

[00:25:29] [SPEAKER_01]: Yeah, it's like a frictionless

[00:25:30] [SPEAKER_01]: way. We need a frictionless

[00:25:32] [SPEAKER_01]: way to work together.

[00:25:34] [SPEAKER_01]: We'll see.

[00:25:36] [SPEAKER_01]: I think of Amazon and

[00:25:37] [SPEAKER_01]: here.

[00:25:37] [SPEAKER_01]: Yeah, glad.

[00:25:39] [SPEAKER_01]: I think of Amazon and the return

[00:25:41] [SPEAKER_01]: process, which has changed a

[00:25:43] [SPEAKER_01]: little bit. They have changed a

[00:25:44] [SPEAKER_01]: little bit. You have to click a

[00:25:45] [SPEAKER_01]: few more buttons, but it is so

[00:25:48] [SPEAKER_01]: frictionless and easy.

[00:25:49] [SPEAKER_01]: You don't like it.

[00:25:50] [SPEAKER_01]: Return. Goodbye.

[00:25:52] [SPEAKER_01]: No questions asked.

[00:25:53] [SPEAKER_01]: I don't need to receive anything.

[00:25:54] [SPEAKER_01]: Well, obviously with this online,

[00:25:55] [SPEAKER_01]: but it's easy

[00:25:57] [SPEAKER_01]: and that makes me go back to

[00:25:59] [SPEAKER_01]: Amazon instead of going to

[00:26:00] [SPEAKER_01]: the store because

[00:26:01] [SPEAKER_02]: you see what I'm see what I'm

[00:26:03] [SPEAKER_02]: struggling with, Ryan is

[00:26:05] [SPEAKER_02]: that's centralized.

[00:26:07] [SPEAKER_02]: The Amazon can represent a

[00:26:08] [SPEAKER_02]: thousand sure.

[00:26:09] [SPEAKER_02]: Yeah, but they still have it

[00:26:11] [SPEAKER_02]: centralized.

[00:26:12] [SPEAKER_02]: Chat GPT for

[00:26:14] [SPEAKER_02]: is going to get to know you

[00:26:15] [SPEAKER_02]: really well.

[00:26:16] [SPEAKER_02]: Yeah, and then all of a sudden

[00:26:18] [SPEAKER_02]: you're going to move over to

[00:26:19] [SPEAKER_02]: Gemini or or you're going to

[00:26:20] [SPEAKER_02]: roll over to another

[00:26:21] [SPEAKER_02]: right large language model,

[00:26:23] [SPEAKER_02]: right? And

[00:26:25] [SPEAKER_02]: will you have to?

[00:26:27] [SPEAKER_02]: I'm assuming start.

[00:26:28] [SPEAKER_02]: You will you have to start

[00:26:30] [SPEAKER_02]: over? Yes.

[00:26:31] [SPEAKER_02]: And I'm thinking about the

[00:26:33] [SPEAKER_02]: portability of OK,

[00:26:35] [SPEAKER_02]: using chat GPT.

[00:26:36] [SPEAKER_02]: And again, this is dumb

[00:26:37] [SPEAKER_02]: examples, but

[00:26:38] [SPEAKER_02]: but using that and all of a

[00:26:40] [SPEAKER_02]: sudden

[00:26:41] [SPEAKER_02]: having to use something new

[00:26:42] [SPEAKER_02]: and starting all the way over.

[00:26:44] [SPEAKER_02]: It's like how

[00:26:45] [SPEAKER_02]: remember we went voice to text

[00:26:48] [SPEAKER_02]: and you'd use a voice to text

[00:26:49] [SPEAKER_02]: thing.

[00:26:50] [SPEAKER_02]: And if you used one of them,

[00:26:52] [SPEAKER_02]: like I use Google

[00:26:53] [SPEAKER_02]: and if you used it

[00:26:54] [SPEAKER_02]: and you stuck with it,

[00:26:55] [SPEAKER_02]: it got it got

[00:26:57] [SPEAKER_02]: OK.

[00:26:58] [SPEAKER_02]: But if you try to different one

[00:27:00] [SPEAKER_02]: man, it was horrible.

[00:27:02] [SPEAKER_02]: So that's my that's one of my

[00:27:04] [SPEAKER_02]: worries is we get siloed

[00:27:06] [SPEAKER_02]: into one of these products.

[00:27:08] [SPEAKER_02]: And then and then now

[00:27:10] [SPEAKER_02]: now it's a battle of products.

[00:27:14] [SPEAKER_00]: It's about exactly exactly.

[00:27:16] [SPEAKER_00]: And that poses another very

[00:27:17] [SPEAKER_00]: interesting question.

[00:27:18] [SPEAKER_00]: I think which we all

[00:27:19] [SPEAKER_00]: wrestle with with all of these

[00:27:21] [SPEAKER_00]: companies and always have

[00:27:23] [SPEAKER_00]: what right do I have

[00:27:25] [SPEAKER_00]: on my data?

[00:27:26] [SPEAKER_00]: The question of consent,

[00:27:28] [SPEAKER_00]: why should I not be able to

[00:27:30] [SPEAKER_00]: hold it?

[00:27:30] [SPEAKER_00]: I can put my phone number.

[00:27:32] [SPEAKER_00]: So I should be able to put

[00:27:34] [SPEAKER_00]: my data which on which

[00:27:36] [SPEAKER_00]: this model got that would be.

[00:27:38] [SPEAKER_02]: And that's a great company to create.

[00:27:40] [SPEAKER_02]: Seriously, I think we just

[00:27:42] [SPEAKER_02]: created a billion dollar company.

[00:27:43] [SPEAKER_02]: Just that alone is that

[00:27:45] [SPEAKER_01]: I'm going to hit stop recording.

[00:27:48] [SPEAKER_01]: Let's get into the back room.

[00:27:49] [SPEAKER_01]: Never get published.

[00:27:50] [SPEAKER_01]: No.

[00:27:51] [SPEAKER_02]: But just think of that technology,

[00:27:53] [SPEAKER_02]: the portability of being able

[00:27:54] [SPEAKER_02]: to have a suitcase

[00:27:55] [SPEAKER_02]: of that large language model

[00:27:57] [SPEAKER_02]: and in taking that

[00:27:58] [SPEAKER_02]: and being able to import that

[00:27:59] [SPEAKER_02]: into a new model

[00:28:01] [SPEAKER_02]: and getting them up to date.

[00:28:04] [SPEAKER_02]: And then you can use a new model.

[00:28:05] [SPEAKER_02]: Yeah, it's fine.

[00:28:07] [SPEAKER_02]: Well, we've seen somebody's

[00:28:08] [SPEAKER_02]: going to create that company.

[00:28:09] [SPEAKER_01]: We've seen that with

[00:28:11] [SPEAKER_01]: what we just had a conversation

[00:28:13] [SPEAKER_01]: recently with Gage.

[00:28:14] [SPEAKER_01]: Same idea.

[00:28:15] [SPEAKER_01]: You get a score

[00:28:16] [SPEAKER_01]: and the candidate can take that

[00:28:18] [SPEAKER_01]: from employer to employers.

[00:28:19] [SPEAKER_01]: Portable.

[00:28:20] [SPEAKER_01]: And so it's your skill set.

[00:28:22] [SPEAKER_01]: It's your references.

[00:28:23] [SPEAKER_01]: It's your performance reviews.

[00:28:24] [SPEAKER_01]: All of that.

[00:28:25] [SPEAKER_01]: Like at some point

[00:28:27] [SPEAKER_01]: will get there.

[00:28:28] [SPEAKER_02]: That'll be easier on the blockchain

[00:28:29] [SPEAKER_02]: because it'll be you can have

[00:28:31] [SPEAKER_02]: it third party validated.

[00:28:32] [SPEAKER_02]: Right.

[00:28:34] [SPEAKER_02]: What do you what do you think?

[00:28:35] [SPEAKER_02]: Because we do have

[00:28:37] [SPEAKER_02]: I think there's one

[00:28:38] [SPEAKER_02]: maybe two companies in HR tech

[00:28:40] [SPEAKER_02]: or work tech

[00:28:41] [SPEAKER_02]: that has third party

[00:28:44] [SPEAKER_02]: audited AI.

[00:28:46] [SPEAKER_02]: Thinking of a company in Australia

[00:28:48] [SPEAKER_02]: called Rijig

[00:28:49] [SPEAKER_02]: and we're talking to the CEO

[00:28:52] [SPEAKER_02]: a while back and she said

[00:28:53] [SPEAKER_02]: when we started,

[00:28:55] [SPEAKER_02]: you know, we didn't know what we didn't know.

[00:28:57] [SPEAKER_02]: So we started down the path

[00:28:58] [SPEAKER_02]: and then we just like,

[00:28:59] [SPEAKER_02]: OK, internally

[00:29:00] [SPEAKER_02]: we're only going to know so much.

[00:29:02] [SPEAKER_02]: So they hired

[00:29:04] [SPEAKER_02]: the university

[00:29:06] [SPEAKER_02]: a university

[00:29:07] [SPEAKER_02]: and they paid them really well

[00:29:09] [SPEAKER_02]: to then just take their all their algorithms

[00:29:12] [SPEAKER_02]: and then say, OK,

[00:29:14] [SPEAKER_02]: tell us what needs to change

[00:29:16] [SPEAKER_02]: because we could do it internally

[00:29:18] [SPEAKER_02]: and we do it internally.

[00:29:19] [SPEAKER_02]: However,

[00:29:20] [SPEAKER_02]: having someone

[00:29:21] [SPEAKER_02]: look at what you're trying to do

[00:29:23] [SPEAKER_02]: again in your case

[00:29:25] [SPEAKER_02]: to no harm

[00:29:26] [SPEAKER_02]: does it actually do that?

[00:29:28] [SPEAKER_02]: Have you all given that any thought?

[00:29:30] [SPEAKER_02]: The idea of hiring a consulting firm

[00:29:33] [SPEAKER_02]: or just somebody that does this for a living

[00:29:35] [SPEAKER_02]: to third party audit

[00:29:37] [SPEAKER_02]: your AI?

[00:29:40] [SPEAKER_00]: Absolutely. Absolutely.

[00:29:42] [SPEAKER_00]: We we are not

[00:29:44] [SPEAKER_00]: we're not doing

[00:29:46] [SPEAKER_00]: right.

[00:29:47] [SPEAKER_00]: AI use cases where we need

[00:29:48] [SPEAKER_02]: right.

[00:29:49] [SPEAKER_00]: That auditing

[00:29:50] [SPEAKER_00]: but that will be definitely at the top of

[00:29:54] [SPEAKER_00]: my list at least

[00:29:56] [SPEAKER_00]: to put a very strong model governance process

[00:29:58] [SPEAKER_00]: in place.

[00:29:59] [SPEAKER_00]: Right.

[00:30:00] [SPEAKER_00]: And it's not one time auditing, right?

[00:30:01] [SPEAKER_00]: It's a continuous auditing

[00:30:03] [SPEAKER_00]: just like you do your

[00:30:04] [SPEAKER_00]: you use

[00:30:05] [SPEAKER_00]: sock to sock to audits socks audits.

[00:30:08] [SPEAKER_00]: It's

[00:30:09] [SPEAKER_00]: it should be in my mind

[00:30:11] [SPEAKER_00]: it should become a similar rigor

[00:30:13] [SPEAKER_00]: when it comes to the AI models as well

[00:30:15] [SPEAKER_00]: where somebody can

[00:30:17] [SPEAKER_00]: certify those things for you

[00:30:19] [SPEAKER_00]: on an ongoing basis.

[00:30:19] [SPEAKER_00]: And yes, you have to give them data

[00:30:21] [SPEAKER_00]: and you have to

[00:30:21] [SPEAKER_00]: do all all the work.

[00:30:23] [SPEAKER_00]: It adds extra work

[00:30:25] [SPEAKER_00]: but given the challenges we have

[00:30:28] [SPEAKER_00]: with AI and Gen AI in general

[00:30:31] [SPEAKER_00]: about the explainability

[00:30:34] [SPEAKER_00]: of the

[00:30:35] [SPEAKER_00]: of the answers which are being given

[00:30:38] [SPEAKER_00]: or how you're using it

[00:30:39] [SPEAKER_00]: in in a narrow set

[00:30:41] [SPEAKER_00]: there is absolutely

[00:30:43] [SPEAKER_00]: case for third party to come in and audit.

[00:30:46] [SPEAKER_01]: indeed we've got

[00:30:47] [SPEAKER_01]: we've gotten questions recently

[00:30:49] [SPEAKER_01]: around technologies role

[00:30:53] [SPEAKER_01]: in

[00:30:54] [SPEAKER_01]: how a company acts.

[00:30:57] [SPEAKER_01]: I'm some thinking specifically around talent acquisition.

[00:31:00] [SPEAKER_01]: And so maybe the software

[00:31:03] [SPEAKER_01]: is leveraging AI to

[00:31:06] [SPEAKER_01]: create rankings or recommendations.

[00:31:09] [SPEAKER_01]: And the company goes forward

[00:31:11] [SPEAKER_01]: and they make a decision

[00:31:12] [SPEAKER_01]: based off of that recommendation, right?

[00:31:15] [SPEAKER_01]: I know you know where this is going.

[00:31:16] [SPEAKER_01]: So the the talent leaders

[00:31:17] [SPEAKER_01]: their question to us has been

[00:31:20] [SPEAKER_01]: I'm trusting the software.

[00:31:22] [SPEAKER_01]: What if that data is bad?

[00:31:25] [SPEAKER_01]: How do I then cover myself

[00:31:28] [SPEAKER_01]: and is my company liable?

[00:31:30] [SPEAKER_01]: What's your kind of take there?

[00:31:35] [SPEAKER_00]: Yep, the company's absolutely liable.

[00:31:37] [SPEAKER_00]: If you're using a software

[00:31:38] [SPEAKER_00]: if you're using a third party software

[00:31:41] [SPEAKER_00]: that third party risk management

[00:31:43] [SPEAKER_00]: and now even more so with

[00:31:45] [SPEAKER_00]: with the lens of what it's going to provide you

[00:31:47] [SPEAKER_00]: is is your responsibility

[00:31:50] [SPEAKER_00]: to make sure.

[00:31:52] [SPEAKER_00]: So I do think that

[00:31:53] [SPEAKER_00]: we if you're going to use it for hiring

[00:31:56] [SPEAKER_00]: if it's work day or whatever

[00:31:57] [SPEAKER_00]: we all know the story which happened there

[00:32:00] [SPEAKER_00]: we are liable if you're going to use it.

[00:32:03] [SPEAKER_00]: And honestly I personally

[00:32:07] [SPEAKER_00]: I believe there are a lot more use cases

[00:32:10] [SPEAKER_00]: for AI which are

[00:32:14] [SPEAKER_00]: quote unquote less risky

[00:32:18] [SPEAKER_00]: to bring into organizations

[00:32:19] [SPEAKER_00]: than using it to rank and stack

[00:32:23] [SPEAKER_00]: resumes and make hiring decisions.

[00:32:25] [SPEAKER_00]: I personally would not feel comfortable with that

[00:32:27] [SPEAKER_00]: because the I don't I don't believe that

[00:32:30] [SPEAKER_00]: the models are quite there yet

[00:32:32] [SPEAKER_00]: where they can pick up the intangibles

[00:32:34] [SPEAKER_00]: and that they're trained on historical bias data.

[00:32:37] [SPEAKER_03]: That's right.

[00:32:38] [SPEAKER_00]: That is a big problem.

[00:32:39] [SPEAKER_00]: So to me that use case

[00:32:42] [SPEAKER_00]: you have to be very very careful.

[00:32:45] [SPEAKER_00]: Maybe it's a narrow use case

[00:32:46] [SPEAKER_00]: where you're hiring just interns

[00:32:47] [SPEAKER_00]: or you're hiring gig workers

[00:32:50] [SPEAKER_00]: who have a very narrow set of skill sets

[00:32:53] [SPEAKER_00]: but not in a wider sense

[00:32:57] [SPEAKER_00]: without a human in the loop.

[00:33:00] [SPEAKER_01]: So our best case is still to train a human

[00:33:03] [SPEAKER_01]: to remove bias.

[00:33:05] [SPEAKER_02]: Well I think some of the cases that we've seen Ryan

[00:33:08] [SPEAKER_02]: is like say in

[00:33:10] [SPEAKER_02]: Storley's case let's say it's a monitoring

[00:33:12] [SPEAKER_02]: so it's not a screen.

[00:33:14] [SPEAKER_02]: It's a computer monitoring employees.

[00:33:17] [SPEAKER_02]: AI can do a whole hell of a lot to monitor

[00:33:21] [SPEAKER_02]: but then kick it to an employee

[00:33:24] [SPEAKER_02]: and say this has been flagged.

[00:33:27] [SPEAKER_02]: Of course the AI is not saying anything.

[00:33:29] [SPEAKER_02]: This has been flagged

[00:33:30] [SPEAKER_02]: and then the employee can then do the investigation

[00:33:33] [SPEAKER_02]: and make sure that that is true not true

[00:33:35] [SPEAKER_02]: etc something that actually needs to be flagged or not

[00:33:38] [SPEAKER_02]: but I think the models right now

[00:33:40] [SPEAKER_02]: that we're working with is more augmentation

[00:33:42] [SPEAKER_02]: where AI goes and does a bunch of things

[00:33:45] [SPEAKER_02]: and doesn't make a decision.

[00:33:47] [SPEAKER_02]: It just gives information to humans

[00:33:51] [SPEAKER_02]: so that they make decisions.

[00:33:52] [SPEAKER_02]: Make the decision.

[00:33:54] [SPEAKER_01]: So question I have then that comes up

[00:33:56] [SPEAKER_01]: and this is I guess for both of you

[00:33:57] [SPEAKER_01]: but if you really

[00:34:01] [SPEAKER_01]: a lot of people say and you say it this way

[00:34:03] [SPEAKER_01]: you say trust but verify.

[00:34:05] [SPEAKER_01]: Right?

[00:34:06] [SPEAKER_01]: So don't spit your water.

[00:34:08] [SPEAKER_01]: Trust but verify.

[00:34:08] [SPEAKER_02]: Reagan actually said that.

[00:34:10] [SPEAKER_02]: Well yes, yes, yes.

[00:34:12] [SPEAKER_01]: So he did yes you're correct.

[00:34:14] [SPEAKER_01]: So okay so scenario

[00:34:16] [SPEAKER_01]: and I'm curious to get your take here.

[00:34:19] [SPEAKER_01]: I'm using an AI.

[00:34:21] [SPEAKER_01]: I'm using a platform to do something right maybe

[00:34:24] [SPEAKER_01]: it's giving me it's not ranking a candidate

[00:34:27] [SPEAKER_01]: or anything like that

[00:34:28] [SPEAKER_01]: but it's giving me information.

[00:34:29] [SPEAKER_01]: It says here Ryan now take this

[00:34:32] [SPEAKER_01]: and you go make your decision.

[00:34:36] [SPEAKER_01]: So now I'm taking that information.

[00:34:38] [SPEAKER_01]: I'm trusting it because my company said yes

[00:34:42] [SPEAKER_01]: this is the software that we trust.

[00:34:44] [SPEAKER_01]: Now I need to go verify.

[00:34:46] [SPEAKER_01]: How do I verify that?

[00:34:48] [SPEAKER_01]: Am I going to another AI to verify it?

[00:34:50] [SPEAKER_01]: Am I going to Google to verify that?

[00:34:53] [SPEAKER_01]: Again, I know it depends on what information

[00:34:55] [SPEAKER_01]: but at the fingertips of a recruiter

[00:34:58] [SPEAKER_01]: or an HR manager or a talent director

[00:35:01] [SPEAKER_01]: how do they verify this properly?

[00:35:06] [SPEAKER_00]: Verify what though

[00:35:07] [SPEAKER_00]: because it's very contextual Ryan.

[00:35:11] [SPEAKER_00]: What are we going to do?

[00:35:12] [SPEAKER_00]: In the case of a background check

[00:35:13] [SPEAKER_02]: it could be a criminal conviction.

[00:35:17] [SPEAKER_02]: So something like that that gets flagged

[00:35:20] [SPEAKER_02]: right then that person can can verify

[00:35:24] [SPEAKER_02]: first of all it should have sources

[00:35:25] [SPEAKER_02]: of where it found that

[00:35:27] [SPEAKER_02]: to where you could go back

[00:35:28] [SPEAKER_02]: and see the source material

[00:35:30] [SPEAKER_02]: to then say okay I got that wrong

[00:35:32] [SPEAKER_02]: or it got that right.

[00:35:34] [SPEAKER_02]: So a person was convicted

[00:35:35] [SPEAKER_02]: but then the case was thrown out on a technicality.

[00:35:39] [SPEAKER_02]: Go to your local municipality whatever is

[00:35:41] [SPEAKER_02]: right whatever the source material is.

[00:35:44] [SPEAKER_00]: Right so the multiple layers are needed

[00:35:46] [SPEAKER_00]: and that's what we do.

[00:35:48] [SPEAKER_00]: So we don't just

[00:35:49] [SPEAKER_00]: in the entire process chain that we have

[00:35:54] [SPEAKER_00]: whether we use AI or not

[00:35:55] [SPEAKER_00]: you get something from one place

[00:35:58] [SPEAKER_00]: but then you go to the actual source

[00:35:59] [SPEAKER_00]: because this is a real case

[00:36:01] [SPEAKER_00]: that there could be a conviction in one code

[00:36:04] [SPEAKER_00]: but then it was thrown out seven years later

[00:36:06] [SPEAKER_00]: and all those kind of cases

[00:36:08] [SPEAKER_00]: come up with or without AI

[00:36:10] [SPEAKER_00]: so you have to go to the source

[00:36:11] [SPEAKER_00]: to do that verification

[00:36:12] [SPEAKER_00]: whether it's manual

[00:36:14] [SPEAKER_00]: whether it's automated

[00:36:16] [SPEAKER_00]: what AI really helps you in those cases to do

[00:36:20] [SPEAKER_00]: you don't have to read

[00:36:22] [SPEAKER_00]: 100 documents.

[00:36:24] [SPEAKER_00]: Which actually you can make more mistakes

[00:36:27] [SPEAKER_00]: reading 100 documents

[00:36:28] [SPEAKER_00]: versus giving it to a machine to say

[00:36:31] [SPEAKER_00]: could you pull this out for me

[00:36:32] [SPEAKER_00]: and find this information

[00:36:35] [SPEAKER_00]: and I actually don't miss information.

[00:36:37] [SPEAKER_00]: That's where

[00:36:38] [SPEAKER_00]: and that is something we have used at Sterling

[00:36:41] [SPEAKER_00]: the power of reading from document

[00:36:44] [SPEAKER_00]: and summarizing it

[00:36:46] [SPEAKER_00]: and actually pulling all the relevant data

[00:36:48] [SPEAKER_00]: versus training humans to go through

[00:36:52] [SPEAKER_00]: 20 code documents or 30 code documents.

[00:36:56] [SPEAKER_00]: Yeah so it can be used for that power of good

[00:37:00] [SPEAKER_00]: if you will in this case

[00:37:01] [SPEAKER_00]: versus just relying on odd brains

[00:37:02] [SPEAKER_00]: to read a ton of information.

[00:37:04] [SPEAKER_02]: Are you building your own LLM for Sterling?

[00:37:08] [SPEAKER_02]: You're using some, okay.

[00:37:09] [SPEAKER_02]: I don't want to get into proprietary stuff

[00:37:11] [SPEAKER_02]: I just was curious about that.

[00:37:14] [SPEAKER_02]: The question I have is

[00:37:17] [SPEAKER_02]: what's keeping you up at night

[00:37:19] [SPEAKER_02]: as it relates to AI?

[00:37:22] [SPEAKER_02]: I'm sure you sleep well so.

[00:37:26] [SPEAKER_00]: What keeps me up at night is

[00:37:29] [SPEAKER_00]: AI attacks,

[00:37:32] [SPEAKER_00]: AI security attacks

[00:37:33] [SPEAKER_00]: because I believe the

[00:37:35] [SPEAKER_00]: the frequency of security attacks

[00:37:37] [SPEAKER_00]: is going to increase

[00:37:39] [SPEAKER_00]: it's already increasing

[00:37:42] [SPEAKER_00]: because of the AI

[00:37:45] [SPEAKER_00]: capabilities and

[00:37:47] [SPEAKER_00]: whether it's a phishing scam

[00:37:48] [SPEAKER_00]: or whether it's

[00:37:49] [SPEAKER_00]: the ability of AI to

[00:37:53] [SPEAKER_00]: run scripts

[00:37:54] [SPEAKER_00]: you know all different kind of things

[00:37:56] [SPEAKER_00]: which can happen.

[00:37:57] [SPEAKER_00]: That's what keeps me up

[00:37:59] [SPEAKER_00]: at night

[00:38:00] [SPEAKER_00]: and I know the

[00:38:01] [SPEAKER_00]: the security tooling is trying to catch up

[00:38:04] [SPEAKER_00]: with it

[00:38:05] [SPEAKER_00]: but the

[00:38:07] [SPEAKER_00]: hacker community is moving very, very fast

[00:38:10] [SPEAKER_00]: and exploiting

[00:38:11] [SPEAKER_00]: yeah that's what keeps me up at night.

[00:38:13] [SPEAKER_02]: If I were in your position

[00:38:15] [SPEAKER_02]: that's exactly what would keep me up at night

[00:38:17] [SPEAKER_02]: is privacy and security.

[00:38:19] [SPEAKER_02]: Just how do I keep people safe?

[00:38:22] [SPEAKER_02]: Getting back to that do not harm

[00:38:24] [SPEAKER_02]: that's both your clients

[00:38:25] [SPEAKER_02]: as well as starving proper.

[00:38:27] [SPEAKER_02]: Yep I got it.

[00:38:28] [SPEAKER_02]: Brian?

[00:38:30] [SPEAKER_01]: You know as you're saying that

[00:38:33] [SPEAKER_01]: I was thinking this is exactly the reason

[00:38:35] [SPEAKER_01]: other than I'm just

[00:38:36] [SPEAKER_01]: I'm not smart enough to do what you do

[00:38:38] [SPEAKER_01]: 100%

[00:38:39] [SPEAKER_01]: not even close

[00:38:40] [SPEAKER_01]: but that is a reason why

[00:38:42] [SPEAKER_01]: I would never be able to do

[00:38:44] [SPEAKER_01]: what you do because I just

[00:38:46] [SPEAKER_01]: I don't think corporate

[00:38:48] [SPEAKER_01]: or say software

[00:38:49] [SPEAKER_01]: however you want platforms

[00:38:50] [SPEAKER_01]: are ever going to outpace the hackers.

[00:38:53] [SPEAKER_01]: It is a constant bad it's never going to

[00:38:57] [SPEAKER_02]: bad actors because there's no rules.

[00:39:00] [SPEAKER_02]: Bad actors can innovate faster.

[00:39:01] [SPEAKER_01]: Yeah and it's not a question of

[00:39:03] [SPEAKER_01]: why can't you think ahead of them

[00:39:06] [SPEAKER_01]: it's I mean why not go higher than

[00:39:08] [SPEAKER_01]: well because then they're already in your system

[00:39:10] [SPEAKER_01]: right like I mean there's a lot of

[00:39:12] [SPEAKER_01]: like that

[00:39:13] [SPEAKER_01]: that not just keeps would keep me up

[00:39:15] [SPEAKER_01]: and I prevent me from ever holding

[00:39:17] [SPEAKER_01]: your job so

[00:39:19] [SPEAKER_01]: I give you a lot of kudos

[00:39:21] [SPEAKER_01]: for that so there's no question in that it was just

[00:39:24] [SPEAKER_00]: thank you thank you

[00:39:25] [SPEAKER_01]: it was just an observation of

[00:39:28] [SPEAKER_01]: you've confirmed that I would never have your role.

[00:39:31] [SPEAKER_01]: What's

[00:39:33] [SPEAKER_00]: I will never be able to do what you're doing.

[00:39:36] [SPEAKER_02]: There you go I feel a little more important.

[00:39:37] [SPEAKER_02]: Last thing for me is

[00:39:39] [SPEAKER_02]: okay so we talked about the dark side

[00:39:40] [SPEAKER_02]: let's talk about the light side

[00:39:41] [SPEAKER_02]: what

[00:39:44] [SPEAKER_02]: what excites you

[00:39:45] [SPEAKER_02]: just immediately

[00:39:47] [SPEAKER_02]: just right around a quarter about AI

[00:39:48] [SPEAKER_02]: like what's this year

[00:39:51] [SPEAKER_02]: you just you're you're giddy about

[00:39:53] [SPEAKER_02]: this phase that we're in in a

[00:40:00] [SPEAKER_00]: you know one thing which I am really hopeful

[00:40:03] [SPEAKER_00]: that which can happen is

[00:40:05] [SPEAKER_00]: and in both personal and professional capacity is

[00:40:10] [SPEAKER_00]: if I don't have to be on phone

[00:40:12] [SPEAKER_00]: with a customer service rep for 40 minutes

[00:40:15] [SPEAKER_00]: in that let's say in a healthcare setting or

[00:40:18] [SPEAKER_00]: if I think of Sterling proper as a candidate

[00:40:21] [SPEAKER_00]: I don't have to be on phone

[00:40:23] [SPEAKER_00]: and I have

[00:40:25] [SPEAKER_00]: I don't know an AI clone

[00:40:26] [SPEAKER_00]: on either side or both sides

[00:40:29] [SPEAKER_00]: who can do that question answer

[00:40:31] [SPEAKER_00]: and then bring me in at the right time

[00:40:35] [SPEAKER_00]: the amount of time it will save

[00:40:36] [SPEAKER_00]: a lot of frustration it will take away

[00:40:40] [SPEAKER_00]: in just

[00:40:41] [SPEAKER_00]: personal professional both sides of life

[00:40:44] [SPEAKER_00]: wouldn't it be amazing

[00:40:45] [SPEAKER_01]: you know when you you're in queue

[00:40:47] [SPEAKER_01]: and it says

[00:40:48] [SPEAKER_01]: if you want to receive a call back

[00:40:51] [SPEAKER_01]: yes I do and then the phone rings

[00:40:53] [SPEAKER_01]: and I get back on the phone and it says

[00:40:55] [SPEAKER_01]: please wait as we find an opera

[00:40:57] [SPEAKER_01]: and just 10 minutes later

[00:40:58] [SPEAKER_01]: wouldn't it be great

[00:40:59] [SPEAKER_01]: for them to get the same message from you

[00:41:01] [SPEAKER_01]: like they're asking your clone

[00:41:03] [SPEAKER_01]: and then your clone says

[00:41:05] [SPEAKER_01]: please hold for if need

[00:41:06] [SPEAKER_01]: she'll be with you momentarily

[00:41:08] [SPEAKER_01]: to answer that question

[00:41:09] [SPEAKER_01]: and 10 minutes later you get on the call

[00:41:11] [SPEAKER_01]: like that would be amazing

[00:41:12] [SPEAKER_02]: I think sinking all that I love that

[00:41:14] [SPEAKER_02]: but the optimism

[00:41:16] [SPEAKER_02]: of that's such an inefficient

[00:41:19] [SPEAKER_02]: process

[00:41:20] [SPEAKER_02]: and it sucks up our time

[00:41:22] [SPEAKER_02]: Ryan chats with a lot of the technologies

[00:41:25] [SPEAKER_02]: that we use for the show

[00:41:26] [SPEAKER_02]: and it's

[00:41:28] [SPEAKER_01]: but those are humans

[00:41:29] [SPEAKER_01]: they're actual humans

[00:41:30] [SPEAKER_01]: they're just bad at their job

[00:41:32] [SPEAKER_01]: I'd rather have an AI at this point

[00:41:35] [SPEAKER_01]: there are some where the AI

[00:41:37] [SPEAKER_01]: is actually better than the humans

[00:41:38] [SPEAKER_01]: all of the platforms that we use

[00:41:40] [SPEAKER_02]: yes

[00:41:42] [SPEAKER_02]: but

[00:41:42] [SPEAKER_02]: you know again it's like

[00:41:45] [SPEAKER_02]: they'll respond in email

[00:41:46] [SPEAKER_02]: this is how inefficient it is

[00:41:48] [SPEAKER_02]: he'll start a chat

[00:41:50] [SPEAKER_02]: you do this too I'm sure

[00:41:51] [SPEAKER_02]: you'll start a chat

[00:41:52] [SPEAKER_02]: and all of a sudden he's

[00:41:54] [SPEAKER_02]: okay enough of this

[00:41:54] [SPEAKER_02]: I've got to get

[00:41:55] [SPEAKER_02]: I've got to move on to the next thing

[00:41:56] [SPEAKER_02]: and then they'll send an email and say

[00:41:58] [SPEAKER_02]: it looks like

[00:41:59] [SPEAKER_02]: you left

[00:42:02] [SPEAKER_02]: and the chat's over

[00:42:03] [SPEAKER_02]: it's like

[00:42:04] [SPEAKER_02]: I left because no one was responding

[00:42:06] [SPEAKER_02]: any way

[00:42:07] [SPEAKER_02]: this has been so educational

[00:42:10] [SPEAKER_02]: thank you so much for Carbon Outtide

[00:42:12] [SPEAKER_02]: we can't even imagine how busy you are

[00:42:14] [SPEAKER_02]: but thanks for coming on the show

[00:42:15] [SPEAKER_02]: breaking things down for us