0:03

Hello, Hello, Welcome to Smart Talks

0:05

with IBM, a podcast from Pushkin

0:07

Industri's iHeartRadio and

0:10

IBM. I'm Malcolm Gladwell.

0:13

This season, we're continuing our conversation with

0:15

new creators visionaries

0:17

who are creatively applying technology in

0:19

business to drive change, but with a

0:21

focus on the transformative

0:24

power of artificial intelligence

0:26

and what it means to leverage AI as

0:28

a game changing multiplier for

0:31

your business. Our guest today

0:33

is Christina Montgomery, IBM's

0:36

Chief Privacy and Trust Officer.

0:38

She's also chair of IBM's AI Ethics

0:41

Board. In addition to overseeing

0:43

IBM's privacy policy, a core

0:45

part of Christina's job involves

0:48

AI governance, making

0:50

sure the way AI is used complies

0:53

with the international legal regulations

0:55

customized for each industry.

0:58

In today's episode, Ristina will explain

1:01

why businesses need foundational

1:03

principles when it comes to using technology,

1:07

why AI regulation should focus

1:09

on specific use cases over

1:11

the technology itself, and share

1:14

a little bit about her landmark congressional

1:17

testimony. Last May, Christina

1:19

spoke with doctor Lori Santos, host

1:22

of the Pushkin podcast The Happiness

1:24

Lab, a cognitive scientist

1:27

and psychology professor at Yale University.

1:29

Laurie is an expert on human

1:32

happiness and cognition. Okay,

1:35

let's get to the interview. So

1:39

Christina, I'm so excited to talk to you today. So

1:42

let's start by talking a little bit about your role

1:44

at IBM. What does a Chief Privacy

1:46

and Trust Officer actually do. It's

1:48

a really dynamic profession and

1:50

it's not a new profession, but the

1:52

role has really changed. I mean, my role

1:55

today is broader than just helping

1:58

to ensure compliance with data protects

2:00

laws globally. I'm also responsible

2:02

for AI governance I co chair or

2:05

AI Ethics Board here at IBM, and

2:07

for data clearance and data governance

2:09

as well for the company. So I

2:12

have both a compliance aspect to my role,

2:14

really important on a global basis, but also

2:17

help the business to competitively

2:19

differentiate because really

2:21

trust is a strategic advantage for IBM

2:24

and a competitive differentiator as a company

2:26

that's been responsibly managing

2:28

the most sensitive data for our clients for

2:30

more than a century now and helping

2:33

to usher new technologies into the world with trust

2:35

and transparency. And so that's also

2:37

a key aspect of my role. And

2:40

so you joined us here on smart talks back

2:42

in twenty twenty one, and you chatted with us about

2:44

IBM's approach of building trust and transparency

2:46

with AI, and that was only two

2:49

years ago. But it almost feels like in eternity

2:51

has happened in the field of AI since then,

2:53

and so I'm curious how much has changed

2:55

since you were here last time. We're the

2:58

things you told us before you are they still

3:00

true? How are things changing? You're

3:02

absolutely right, it feels like the world

3:05

has changed really in the last two

3:07

years. But the same fundamental

3:09

principles and the same overall governance

3:12

apply to IBM's program

3:15

for data protection and responsible

3:17

AI that we talked about two years

3:20

ago, and not much has changed there from

3:22

our perspective. And the good thing

3:24

is we've put these practices and

3:26

this governance approach into place,

3:29

and we have an established way

3:31

of looking at these emerging technologies. As

3:33

the technology evolves, the tech

3:35

is more powerful, for sure, foundation models

3:38

are vastly larger and more capable,

3:41

and our creating in some respects new

3:43

issues. But that just makes it all the

3:45

more urgent to do what we've been doing and

3:47

to put trust and transparency into place

3:49

across the business to be accountable

3:52

to those principles, and

3:54

so our conversation today is really centered around this

3:56

need for new AI regulation and

3:58

part of that regulations the mitigation

4:01

of bias. And this is something I think about

4:03

a ton as a psychologist, right, you know, I

4:05

know, like my students and everyone who's

4:07

interacting with AI is assuming

4:09

that the kind of knowledge that they're getting

4:11

from this kind of learning is accurate,

4:13

right, But of course AI is only as good as the knowledge

4:16

that's going in. And so talk to me a

4:18

little bit about like why bias

4:20

occurs in AI and the level of the problem

4:22

that we're really dealing with. Yeah,

4:24

Well, obviously AI is based on data,

4:27

rights it's trained with data,

4:30

and that data could be biased

4:32

in and of itself, and that's where issues

4:35

could come up. They come up in the data, they could

4:37

also come up in the output of the models

4:39

themselves. So it's really

4:41

important that you build bias

4:44

consideration and bias testing into

4:46

your product development cycle. And

4:48

so what we've been thinking about here at

4:50

IBM and doing we had some of our research

4:52

teams delivered some of the very first

4:55

toolkits to help detect bias years

4:57

ago now, right, and deployed them to open source,

5:00

and we have put into place for

5:02

our developers here at IBM and

5:04

Ethics by Design playbook that's

5:07

sort of a step by step approach which

5:09

also addresses very fully

5:12

biased considerations, and

5:14

we provide not only like

5:17

here's a point when you should test for it and

5:20

you consider it in the data, you have to measure

5:22

it both at the data and the model level or the

5:24

outcome level, and we provide

5:26

guidance with respect to what tools

5:28

can best be used to accomplish that. So

5:31

it's a really important issue. It's one

5:33

you can't just talk about. You have to provide

5:35

essentially the technology and the capabilities

5:38

and the guidance to enable people to test

5:40

for it. Recently, you had this wonderful

5:42

opportunity to head to Congress to talk about

5:44

AI, and in your testimony before

5:46

Congress you mentioned that it's often said

5:49

that innovation moves too fast for government

5:51

to keep up. And this is something that I also

5:53

worry about as a psychologist. Right our policy

5:55

makers really understanding the issues that they're

5:57

dealing with, and so I'm curious how you're approaching

5:59

the challenge of adapting AI policies

6:02

to keep up with the sort of rapid pace of all

6:04

the advancements we're seeing in the AI technology

6:06

itself. I think it's really

6:09

critically important that you have foundational

6:11

principles that applied to not

6:14

only how you use technology,

6:16

but whether you're going to use it in the first place, and where

6:19

you're going to use and apply it across your company.

6:21

And then your program from a governance perspective,

6:24

has to be agile. It has to be able

6:26

to address emerging capabilities,

6:29

new training methods, etc. And

6:32

part of that involves helping

6:34

to educate and instill and empower

6:37

a trustworthy culture at a company

6:39

so you can spot those issues so you can

6:41

ask the right questions at the right time if

6:44

you try. We talked about during

6:46

the Senate hearing, and IBM's been talking

6:48

for years about regulating

6:51

the use, not the technology itself,

6:53

because if you try to regulate technology,

6:56

you're very quickly going to find out regulation

6:59

will appbsolutely never keep up

7:01

with that. And so in your testimony to Congress,

7:03

you also talked about this idea of a precision

7:05

regulation approach for AI. Tell

7:08

me more about this. What is a precision regulation

7:10

approach and why could that be so important.

7:12

It's funny because I was able to share with

7:14

Congress our precision

7:16

regulation point of view in

7:18

twenty twenty three, but that precision

7:21

regulation point of view was published by IBM

7:23

in twenty twenty So

7:26

we have not changed our position

7:29

that you should apply the tightest

7:31

controls, the strictest regulatory requirements

7:34

to the technology where the end use

7:37

and risk of societal harm is the greatest.

7:39

So that's essentially what it is. There's

7:42

lots of AI technology that's used

7:44

today that doesn't touch people, that's

7:46

very low risk in nature. And

7:48

even when you think about AI

7:51

that delivers a movie recommendation versus

7:53

AI that is used to diagnose

7:56

cancer, right, there's very different

7:58

implications associated with those two uses

8:00

of the technology. And

8:03

so essentially what precision regulation

8:05

is is apply different rules to different risks,

8:07

right, more stringent regulation to the use

8:10

cases with the greatest risk. And

8:12

then also we build that

8:14

out calling for things like transparency

8:17

you see it today with content right,

8:19

misinformation and the like. We

8:22

believe that consumers should always

8:24

know when they're interacting with an AI system,

8:26

So be transparent, don't hydra AI clearly

8:30

define the risks. So as a

8:32

country, we need to have some clear guidance

8:34

right in globally, as well in

8:36

terms of which uses

8:39

of AI or higher risk where will apply

8:41

higher and stricter regulation and

8:44

have sort of a common understanding of what

8:46

those high risk uses are and

8:49

then demonstrate the impact in the cases

8:51

of those higher risk uses.

8:53

So companies who are

8:55

using AI in spaces where they can impact

8:58

people's legal rights, example,

9:01

should have to conduct an impact

9:03

assessment that demonstrates that

9:05

the technology isn't biased. So we've

9:07

been pretty clear about apply that

9:10

the most stringent regulation to the highest

9:12

risk uses of AI. And

9:15

so so far we've been talking about your congressional

9:17

testimony in terms of, you know, the specific

9:19

content that you talked about, But I'm just curious on

9:21

a personal level, you know, what was that

9:24

like right like right now it feels like at a policy

9:26

level, like there's a kind of fever pitch going

9:28

on with AI right now. You know what did that

9:30

feel like to kind of really have the opportunity

9:32

to talk to policy makers and sort of influence

9:34

what they're thinking about AI technologies like

9:37

in the coming century. Perhaps I

9:39

was really an honor to be able to do

9:41

that and to be one

9:43

of the first set of invitees to

9:45

the first hearing, and what I

9:48

learned from it essentially is you know,

9:50

really two things. The first is really

9:52

the value of authenticity. So

9:54

both as an individual and

9:56

as a company, I was

9:59

able to talk about out what I do.

10:01

You know, I need a lot of advanced prep

10:03

right. I talked about what my

10:05

job is, what IBM

10:07

has been putting in place for years now. So

10:11

this isn't about creating something. This

10:13

was just about showing up and being authentic.

10:15

And we were invited for a reason. We were invited

10:18

because we were one of the earliest

10:20

companies in the AI technology

10:22

space. We're the oldest

10:24

technology company and we

10:27

are trusted and that's an

10:29

honor. And then the second thing I came

10:31

away with was really how important this issue is

10:33

to society. I don't think I appreciated it

10:36

as much until following

10:38

that experience. I had

10:41

outreached from colleagues I hadn't worked with

10:43

for years. I had an outreach from family

10:45

members who heard me on the radio. You

10:48

know, my mother and my mother in law,

10:50

and my nieces and nephews and my friends

10:52

of my kids were all like, oh, I get it, I

10:54

get what you do. Now, Wow, that's pretty cool,

10:57

you know. So that was really probably

10:59

the best, the most impactful takeaway

11:01

that I had. The mass adoption of

11:03

generative AI happening at breakneck

11:06

speed has spurred societies

11:08

and governments around the world to

11:10

get serious about regulating

11:13

AI. For businesses,

11:15

compliance is complex enough already, but

11:17

throw an ever involving technology like AI

11:20

into the mix, and compliance itself

11:22

becomes an exercise in adaptability.

11:26

As regulators seek greater accountability

11:29

in how AI is used, businesses

11:31

need help creating governance processes

11:34

comprehensive enough to comply with

11:36

the law, but agile enough to

11:39

keep up with the rapid rate of change in

11:41

AI development. Regulatory

11:43

scrutiny isn't the only consideration

11:46

either responsible AI governance.

11:49

Of business's ability to prove its AI

11:51

models are transparent and explainable

11:55

is also key to building trust with customers,

11:58

regardless of industry. In

12:00

the next part of their conversation, Laurie

12:03

asked Christina what businesses should

12:05

consider when approaching AI

12:07

governance. Let's listen. So

12:10

it's a particular role that businesses are playing

12:12

in AI governance, Like, why is it so critical

12:15

for businesses to be part of this? So

12:17

I think it's really critically important

12:19

that businesses understand

12:22

the impacts that technology can have both

12:24

in making them better businesses, but

12:27

the impacts that those technologies can have on

12:30

the consumers that they

12:32

are supporting. You know, businesses

12:34

need to be deploying AI

12:37

technology that is in alignment

12:39

with the goals that they set for it, and that can

12:41

be trusted. I think for us and for our

12:43

clients, a lot of this comes back

12:46

to trust in tech. If

12:48

you deploy something that doesn't work, that

12:50

hallucinates, that discriminates,

12:54

that isn't transparent, where decisions

12:56

can't be explained, then you are

12:58

going to very rapidly erode

13:00

the trust at best right of your

13:03

clients and at worst for yourself.

13:05

You're going to create legal and regulatory issues

13:07

for yourself as well. So trusted technology

13:10

is really important, and I think there's

13:12

a lot of pressure on businesses today to move very

13:14

rapidly and adopt technology. But if you

13:16

do it without having a program of governance

13:19

in place, you're really risking eroding

13:21

that trust. So this is really where I think a

13:23

strong AI governance comes in. Talk

13:26

about from your perspective, how this really

13:28

contributes to maintaining the trust

13:30

that customers and stakeholders have in these

13:32

technologies. Yeah, absolutely,

13:34

I mean you need to have a governance program

13:36

because you need to understand that

13:39

the technology, particularly in the AI space,

13:41

that you are deploying, is

13:44

explainable. You need to understand

13:46

why it's making decisions and

13:48

recommendations that it's making, and you need to be able

13:51

to explain that to your consumers. I mean, you

13:53

can't do that if you don't know where your data is

13:55

coming from, what data are you using to train those

13:57

models if you don't have a program

13:59

that manage is the alignment

14:01

of your AI models over time to make

14:04

sure as AI learns

14:06

and evolves over uses, which

14:08

is in large part what

14:11

makes it so beneficial that

14:13

it stays in alignment with the objectives

14:15

that you set for the technology over

14:17

time. So you can't

14:19

do that without a robust governance

14:22

process in place. So we work with

14:24

clients to share our own

14:26

story here at IBM in terms of how

14:28

we put that in place, but also in

14:30

our consulting practice to

14:33

help clients work with these

14:35

new generative capabilities and foundation

14:38

models and the like in order to put

14:40

them to work for their business in a way that's going

14:42

to be impactful to that business but

14:44

at the same time be trusted. And so

14:46

now I wanted to turn a little bit towards watsonex

14:49

governance, and so IBM recently announced

14:51

their AI platform Watson X, which

14:53

will include a governance component. Could

14:56

you tell us a little more about watsonex dot

14:58

Governance. Yeah, I mean before do that,

15:00

I'll just back up and talk about the full platform

15:04

and then lean into Watson X because I

15:06

think it's important to understand the delivery

15:09

of a full suite of capabilities

15:12

to get data, to train

15:15

models, and then to govern them over their

15:17

life cycle. All of

15:19

these things are really important

15:22

from the onset. You need to make

15:24

sure that you have, you know, for our

15:26

watsonex dot AI for example,

15:30

that's the studio to train new foundation

15:32

models and generative

15:34

AI and machine learning capabilities.

15:37

And we are populating

15:40

that studio with some IBM

15:42

trained foundation models which we're

15:45

curating and tailoring

15:47

more specifically for enterprises. So

15:49

that's really important. It comes back to the point I made

15:51

earlier about business trust and

15:54

the need you know, to have

15:56

enterprise ready technologies

15:59

in the AI space. And then the

16:01

watsonex dot data is a

16:03

fit for purpose data store or a data

16:06

lake. And then watsonex dot

16:08

gov so that's a particular

16:10

component of the platform that

16:13

my team and the AI Ethics Board

16:16

has really worked closely with the product

16:18

team on developing, and we're using

16:20

it internally here in the Chief Privacy

16:22

Office as well to help us

16:25

govern our own uses of AI

16:27

technology and our compliance

16:30

program here, and it essentially

16:33

helps to notify you

16:35

if a model becomes biased or gets

16:38

out of alignment as you're using it over

16:40

time. So companies are going to need

16:42

these capabilities. I mean they need them today

16:44

to deliver technologies with

16:47

trust. They'll need them tomorrow

16:49

to comply with regulation, which is on

16:51

the horizon, and I think compliance becomes

16:53

even more complex when you consider international

16:56

data protection laws and regulations. Honestly,

16:59

I don't know how any on any company's legal

17:01

teams keeping up with this these days. But my

17:03

question for you is really how can businesses

17:06

develop a strategy to maintain

17:08

compliance and to deal with it in this ever changing

17:10

landscape. It's increasingly more challenging.

17:12

In fact, I saw statistic just

17:14

this morning that the regulatory

17:17

obligations on companies have increased something

17:19

like seven hundred times in

17:22

the last twenty years, So it really

17:24

is a huge focus

17:26

area for companies. You have to have a

17:28

process in place in order

17:30

to do that, and it's not easy, particularly

17:33

for a company like IBM that

17:35

it has a presence in over one hundred

17:37

and seventy countries around the world. There

17:40

is more than one hundred and fifty comprehensive

17:43

privacy regulations, there

17:45

are regulations of non personal

17:47

data, there are AI regulations emerging,

17:50

so you really need an operational approach

17:53

to it in order

17:55

to stay compliant. But one of the things we do is

17:57

we set up baseline and a lot of companies

17:59

do that as well. So we define a privacy

18:02

baseline, we define an AI baseline,

18:05

and we ensure then

18:07

as a result of that that there are very few deviances

18:10

because it incorporates in that baseline.

18:12

So that's one of the ways we do it. Other companies,

18:14

I think are similarly situated in

18:17

terms of doing that. But

18:20

again, it is a real challenge

18:22

for global companies. It's one of the reasons why

18:25

we advocate for as much alignment as

18:27

possible on the international

18:30

realm as well as nationally

18:32

here in the US, as much alignment

18:35

as possible to make compliance

18:38

easier for easier and not

18:40

just because companies want an easy way

18:42

to comply, but the harder

18:44

it is, the less likely there will

18:46

be compliance. And it's

18:49

not the objective of anybody

18:51

governments, companies consumers

18:54

to have to set legal obligations

18:57

that companies simply can't meet. So

18:59

what advice would you give to other companies who are

19:01

looking to rethink or strengthen their approach

19:03

to AI governance. Think you need to start with,

19:05

as we did, foundational principles,

19:08

and you need to start making decisions

19:10

about what technology you're going to deploy

19:13

and what technology you're not, What are you going to use it for, and

19:15

what aren't you going to use it for? And then when you do use

19:17

it, align to those principles.

19:20

That's really important. Formalize a program,

19:22

have someone within the organization,

19:25

whether it's the chief privacy officer, whether

19:28

it's some other role, a chief

19:30

AI ethics officer, but have

19:32

an accountable individual and accountable

19:35

organization. Do a maturity

19:37

assessment, figure out where you are and where you need

19:39

to be, and really start

19:42

putting it into place today. Don't

19:45

wait for regulation to apply

19:47

directly to your business, because it'll be too

19:49

late. So Smart

19:51

Talks features new creators, these visionaries

19:54

like yourself, who are creatively applying technology

19:56

in business. To drive change. I'm curious

19:59

if you see yourself as creative, you

20:01

know, I definitely do. I

20:03

mean, you need to be creative

20:06

when you're working in an industry

20:08

that evolves so very quickly. So

20:11

you know, I started with IBM

20:14

when we were primarily a hardware company,

20:16

right, and we've changed our business so

20:19

significantly over the years, and the issues

20:21

that are raised with respect to each new

20:24

technology, whether it be cloud,

20:26

whether it be AI now where

20:29

we're seeing a ton of issues, or you look at emergent

20:31

issues in the space of things

20:33

like neurotechnologies and quantum computers.

20:36

You have to

20:38

be strategic and

20:40

you have to be creative and thinking about

20:43

how you can adapt agilely

20:46

quickly a company

20:48

to an environment that is changing

20:50

so quickly and with

20:52

this transformation happening at such a rapid

20:55

pace. Do you think creativity plays a role

20:57

in how you think about and implement specifically

20:59

a trust where the AI strategy.

21:03

Yeah, I absolutely

21:05

think it does, because again, it comes back

21:07

to these capabilities, and there are ways.

21:10

I guess how you define creativity

21:12

could be different, right, But

21:14

I'm thinking of creativity in the sense of

21:17

sort of agility and strategic vision

21:19

and creative problem solving. I

21:21

think that's really important

21:23

in the world that we're in right now, being able

21:26

to creatively problem solve

21:28

with new issues that are rising

21:31

sort of every day. And

21:33

so, how do you see the role of chief privacy officer

21:36

evolving in the future as AI technology

21:38

continues to advance, Like what stuff

21:40

should CPOs take to stay ahead of all these changes

21:43

that are come in their way? So

21:45

the role is evolving in

21:47

most companies, I would say pretty

21:50

rapidly. Many companies

21:52

are looking to chief privacy officers who

21:54

are ready understand the data

21:57

that's being used in the organization and have programs

21:59

to ensure compliance with laws

22:02

that require you to manage that data in

22:04

accordance with data protection laws and the like. It's

22:07

a natural place and position for

22:11

AI responsibility. And

22:13

so I think what's happening to a lot of chief

22:15

privacy officers is they're being asked to

22:17

take on this AI governance responsibility

22:20

for companies and if not take

22:22

it on, at least play a very

22:24

key role working with other parts

22:26

of the business in AI governance.

22:29

So that really is changing. And if chief

22:32

privacy officers are in companies

22:34

who maybe haven't started thinking about AI

22:36

yet, they should so I would

22:39

encourage them to look at

22:41

different resources that are available already

22:43

in AI governance space. For

22:45

example, the International Association of

22:47

Privacy Professionals, which is the

22:50

seventy five thousand member professional

22:53

body for the profession

22:55

of chief Privacy officers, just recently

22:57

launched an AI Governance

23:00

in an AI Governance certification program.

23:03

I sit on their advisory board. But that's

23:05

just emblematic of the fact that the field

23:08

is changing so rapidly. And

23:11

so, you know, speaking of rapid change, when

23:13

you're back here on smart Talks in twenty

23:15

twenty one, you said that the future of AI

23:17

will be more transparent and more trustworthy.

23:20

You know, what do you see the next five to ten years

23:22

holding. You know, when you're back on smart Talks

23:24

in you know, twenty twenty six, you know

23:26

twenty thirty, you know what are we going to be talking about

23:28

when it comes to AI technology and governance.

23:31

So I try to be an optimist, right And

23:33

I said that two years ago, and

23:36

I think we're seeing it now come

23:38

into fruition, and there will be

23:41

requirements, whether they're

23:43

coming from the US, whether they're coming from Europe,

23:45

whether they're just coming from voluntary adoption

23:48

by clients of things like the

23:50

NISS Risk Management Framework, really

23:52

important voluntary frameworks. You're

23:54

going to have to adopt transparent and explainable

23:57

practices in your uses of AI. So

24:00

I do see that happening. And in the next five to ten

24:02

years, boy, I think we'll see more

24:04

research into trust

24:07

in techniques because

24:09

we don't really know, for

24:11

example, how to water mark. We

24:14

were calling for things like watermarking. There'll

24:16

be more research into how to do that. I

24:19

think you'll see regulation

24:23

that's specifically going to require those types of things.

24:26

So I think again, I think the regulation is

24:28

going to drive research. It's going to drive

24:30

research into these areas that will

24:32

help ensure that we can deliver

24:35

new capabilities, generated capabilities

24:37

and the like with trust and explainability.

24:40

Thank you so much Christina for joining me on smart

24:42

Talks to talk about AI and governance. Well,

24:45

thank you very much for having me to

24:48

unlock the transformative growth possible

24:50

with artificial intelligence. Businesses

24:53

need to know what they wish to

24:55

grow into first. Like

24:57

Christina said, the best way forward in

24:59

the AI future is for businesses

25:02

to figure out their own foundational principles

25:05

around using the technology. Drawing

25:07

on those principles, to apply AI

25:09

in a way that's ethically consistent

25:12

with their mission and complies with the legal

25:14

frameworks built to hold the technology

25:17

accountable. As AI

25:19

adoption grows more and more widespread,

25:21

so too will the expectation from

25:23

consumers and regulators that

25:26

businesses use it responsibly. Investing

25:29

independable AI governance is a

25:31

way for businesses to lay the foundations

25:34

for technology that their customers

25:36

can trust while rising to

25:38

the challenge of increasing regulatory

25:41

complexity. Though the emergence

25:43

of AI does complicate an already

25:46

tough compliance landscape, businesses

25:48

now face a creative opportunity

25:51

to set a precedent for what accountability

25:54

in AI looks like and rethink

25:56

what it means to deploy trustworthy

25:59

artificial intelligence. I'm

26:01

Malcolm Gladwell. This is a paid

26:04

advertisement from IBM.

26:07

Smart Talks with IBM will be taking a short hiatus,

26:09

but look for new episodes in the

26:11

coming weeks. Smart Talks

26:14

with IBM is produced by Matt Romano, David

26:17

jaw Nische, Venkat and

26:19

Royston Deserve with Jacob Goldstein.

26:21

We're edited by Lydia Jene Kott.

26:24

Our engineer is Jason Gambrel. Theme

26:26

song by Gramoscope Special

26:29

thanks to Carli Migliori, Andy Kelly,

26:31

Kathy Callahan, and the eight

26:33

Bar and IBM teams, as

26:35

well as the Pushkin Marketing team.

26:38

Smart Talks with IBM is a production

26:40

of Pushkin Industries and Ruby Studio

26:43

at iHeartMedia. To find

26:45

more Pushkin podcasts, listen on

26:48

the iHeartRadio app, Apple Podcasts,

26:50

or wherever you listen to

26:53

podcasts.