2:00

and migration from rural to urban areas.

2:03

Now, in the face of artificial intelligence,

2:05

the question many are grappling with is

2:07

whether machines will replace us. But

2:10

the more pressing issue may not be

2:12

that these systems are replacing human labor,

2:14

but rather concealing it. Take

2:18

the mechanical Turk, built in 1770. It

2:21

was touted as an automaton that could

2:23

play chess, and it made the rounds

2:25

facing off against the likes of Benjamin

2:27

Franklin and Napoleon. But

2:30

the technology behind this machine was

2:32

all a charade. It

2:35

was presented to the public, and

2:37

it toured for decades, and it

2:39

delighted crowds because it looked like

2:42

it was a fully operating clockwork

2:45

automaton that could play

2:47

chess. Now, as many

2:49

of us now know, there was a

2:51

person inside who was hidden in a

2:53

secret compartment. This is tech

2:55

critic and writer Joanne McNeil. She

2:58

says the story of the mechanical Turk over

3:00

two centuries ago is in some

3:02

ways also that of automation today. I

3:06

look around and I can't think of one instance

3:08

of AI where

3:10

you don't see human labor somewhere

3:12

alienated from the actual product in

3:14

the service. Like I

3:16

said, Joanne is a tech critic, but she

3:19

also recently came out with her debut novel.

3:21

It's called Wrong Way, and though it's fiction,

3:23

it has a lot to say about our

3:25

current moment of tech-fueled gig work. It

3:28

centers on a middle-aged woman named Teresa who's

3:30

stuck in a cycle of precarious work. She

3:33

takes a mysterious yet promising job at a

3:35

tech behemoth called All Over.

3:38

I was really interested in someone who

3:41

had maybe what we might

3:43

consider a traditional working class job

3:45

history emerging in the

3:48

2020s into

3:50

the gig economy that Silicon

3:52

Valley has sort of turned

3:56

so many of these once maybe less...

10:00

and kind of like a little office

10:02

space and watching the role of the remote

10:04

operator is kind of like taking

10:06

over for navigation and for

10:08

as few cars as Waymo actually has

10:10

on the road. It's a huge labor

10:12

force of remote operators. So

10:14

that was very, very

10:16

interesting experience for me.

10:19

You are listening to

10:21

Spark. This is Spark.

10:23

This is Spark. From

10:25

CBC. I'm

10:38

Nora Young and today we're talking about Work

10:40

Part 6 in our occasional series Being Human

10:42

Now. Right now my guest

10:45

is Joanne McNeil. Joanne is a tech

10:47

critic and now novelist. Her debut novel

10:49

is called Wrong Way. It explores the

10:51

often hidden human cost of automation and

10:53

gig work. In

10:56

your novel, the CEO of Oliver, who

10:59

has the great name Falconer, he uses

11:01

this very progressive sort of techno-solutionist

11:03

language throughout the book that really

11:05

does feel a lot like the

11:07

manifestos written by Silicon

11:10

Valley giants. Their ethos

11:12

is something called holistic apex.

11:15

So tell me a bit about what you're doing

11:17

with this character. I wanted to show how little

11:20

commitment Silicon Valley leader would

11:22

need to express in their statements.

11:26

And so I thought it

11:28

would be very funny to have a Silicon

11:30

Valley billionaire who's not just necessarily

11:33

claiming he's progressive, but

11:35

claiming he's anti-hierarchical, while

11:37

also giving his money away.

11:39

And it's funny because

11:41

when I have interacted with

11:43

people who are very deep

11:45

in the Silicon Valley mindset, I

11:48

always end up very frustrated. I can't have

11:50

a proper debate with them because they will

11:52

say anything to win. And so when I

11:54

was thinking about writing in his voice, it

11:58

was always the kind of the statement. that

12:00

are dead ends, that are very difficult

12:02

to argue with, like the rhetoric

12:04

that just almost shuts the conversation down, that he

12:06

will say something like, well, antitrust

12:09

is pro-capitalist. Right. But

12:13

is there a power in kind of claiming

12:15

a vision of the future that you then

12:17

kind of own and can sell? Yes,

12:20

because a vision of the

12:22

future is different from the

12:24

vast unknown, and the unknown

12:26

frightens us, uncertainty frightens anyone.

12:29

And when you have these

12:32

kind of leaders who offer

12:34

roadmaps to somewhere,

12:38

that's something you at least can see.

12:40

And in fact, like the other problem

12:42

with that is you're arguing

12:44

with their terms, because if someone were

12:46

to say, you know, Mars

12:49

in five years, that's it, we're doing it,

12:51

we are going to have life on Mars

12:53

in five years. The

12:55

next five years, you have

12:58

to argue with why maybe that's not

13:00

the best use of resources,

13:03

because otherwise, there is this unknown. So I

13:05

think it's a power of science fiction, too.

13:07

And it's a power that I'm always thinking

13:10

through when I write science fiction is

13:12

that once you imagine

13:15

something, it's a

13:17

possibility. It's no longer this

13:19

vague space. With

13:22

my writing, I hope it stands

13:24

out as fiction first. I mean, that is what I intended

13:27

to do. But I also am

13:29

thinking about the future. And one thing I

13:31

wanted to express in this book was that it's

13:34

set in the future, but it's set in a future

13:36

that feels very much like now. It's

13:39

set in the future of a moment

13:41

of decline that in fact, the

13:44

technologies are not getting better. The

13:46

technologies we have today are kind of

13:49

crumbling. The infrastructure is crumbling.

13:53

And this, to me, is the

13:57

reality of the future that

13:59

many... progressive gains that we

14:01

might have seen in our

14:03

recent lifetime, acceptance of trans

14:06

people over the past decade,

14:08

but now in recent

14:11

years, the policy that is

14:14

clawing back that progressive action,

14:16

that we can't necessarily see

14:18

the future as constant wins,

14:20

constant life getting better, but

14:23

as just ongoing change,

14:26

including decline. You

14:30

started writing the wrong way in 2018 and

14:32

it came out at the end of 2023. 2023 was sort of dubbed

14:36

as the year of artificial intelligence. Obviously a

14:38

lot changed in those five year period while

14:41

you were writing it. Could you reflect a little bit

14:43

on the changes that we've seen in that time, especially

14:45

when it comes to AI? Yeah,

14:48

it's an example of

14:51

how swiftly technology can

14:53

be normalized. I want

14:55

to say it was winter of 22

14:58

that people were even hearing about open

15:00

AI for the first time or discovering

15:03

what something like mid journey or

15:05

chat TPT can do that there

15:08

was this very clear moment that

15:10

AI all of a sudden became

15:12

a very mainstream conversation. While people

15:14

might have followed its developments over

15:16

time, I don't think many

15:18

of my friends who don't follow technology would

15:20

have known who Sam Altman was. And then

15:23

all of a sudden, he's

15:25

everywhere. And a year

15:27

isn't enough time to discover

15:31

what is a reasonable

15:33

way to integrate this

15:35

technology into our lives

15:37

if we want to at all. And one

15:40

company is wealthy and powerful.

15:43

And the descent on the other hand

15:45

is scrappy and without

15:47

resources. Something that does

15:49

give me hope is that the descent

15:52

is broad. I think a lot of

15:54

people do have that visceral response to

15:56

LLMs that I described before that it

15:59

feels like like you're going through my

16:01

stuff. Like why do you need my life

16:04

to be crumbled up like that and

16:06

shoot out? Is there some way we

16:08

could do this without that trade that

16:10

isn't a trade? It's just taking

16:12

it. Yeah. So

16:15

what would you like to see included in

16:17

the conversation about, you

16:19

know, automation, gig work, and

16:21

the human labor at the heart of some of

16:23

these automated systems that we use? What are what

16:26

are we not talking about that we need to

16:28

talk about? I'd like to

16:30

see more transparency. I mean, the fact

16:32

that they're kind of very hazy about

16:34

what the training data even is that's

16:36

powering various LMS in

16:39

the case of open AI

16:41

transparency about the workers themselves.

16:43

Why are the remote operators

16:45

so hidden from the public?

16:47

I mean, if they're

16:49

integral to the operation

16:51

of a Waymo vehicle, then

16:54

their role should be much more transparent

16:56

to the public. And we

16:58

see this again and again, what the

17:00

company is hiding from us is not

17:02

just the secret sauce, it's exploitation that

17:05

makes the technology possible. And if a

17:07

technology is based on human exploitation, then

17:09

can we safely agree

17:11

that it shouldn't exist? You've

17:15

written about content, moderation and the conditions of

17:17

their work in particular in this regard. Yeah,

17:19

I always notice when those stories come up

17:21

because I get to a personal place, I

17:23

think that like when I was in my

17:25

20s and trying to find a job, that's

17:27

the kind of job that I might have

17:29

landed on. And so when I

17:31

hear about these traumatized workers

17:34

who might have just thought that they

17:36

were stumbling on something to

17:38

do in the daytime and work

17:40

on music or comic books or

17:42

whatever their hobbies or passions

17:44

are at night, and

17:46

to have not just a terrible office

17:48

job, but an office job that exposes

17:51

you to horrors again

17:53

and again, and really

17:55

pushes you beyond that limit. It's

17:57

so confounding. And I feel known

44:00

as scientific management or

44:02

social engineering. When Friedrich Taylor

44:04

working with some of the

44:06

early assembly lines and factories

44:08

in the US and North

44:10

America more broadly really started

44:13

to look at humans as

44:15

small components of a ginormous

44:17

machine or cockpit and everything

44:19

started to become a vehicle

44:21

or tool in the interest

44:23

of productivity. And in

44:25

the five or six decades that followed,

44:28

actually, there was a big movement towards

44:30

empowering employees, looking after workers rights and

44:32

with that came a lot of good

44:34

legislation and regulation. As

44:37

talent management started to become

44:39

really prominent in the 1990s,

44:41

we actually entered a spiritual

44:43

age where things like employee

44:45

engagement and thriving and career

44:48

fit and talent and potential

44:50

all became really important competencies.

44:52

But with the rise of

44:54

big data and data analytics,

44:56

which includes AI and AI

44:58

surveillance, actually what you have

45:01

is both things operating. On the one

45:03

hand, employers all try to provide employees

45:05

with a sense of purpose, ensure that

45:07

they can thrive and experience calling and

45:09

we hear employees and leaders saying they

45:11

want employees to bring their whole self

45:13

to work and to be themselves and

45:15

that they're valuable for their unique characteristics.

45:18

But on the other hand, underneath it,

45:20

we're monitoring and measuring everything. I

45:22

think there's still this idea that people

45:25

are productivity machines and that if you

45:27

measure the performance and you incentivize them

45:29

and either create nudges, actually they will

45:31

deliver. And so even when we seem

45:34

to care about engagement and happiness, actually

45:36

the ultimate goal is to squeeze as

45:38

much profitability and productivity of workers. And

45:41

when you add to this layer the

45:43

fact that people are so dependent on

45:46

technology and interacting with AI and other

45:48

technologies so much, there's a real need,

45:50

I think, for organizations and leaders to

45:52

rehumanize work and actually rediscover some of

45:55

the things that actually made work interesting

45:57

and valuable in the first place. Some

46:00

critics have even argued that the wellness

46:02

movement in the workplace itself is sort

46:05

of inherently tied to this kind of

46:07

level of efficiency and productivity and surveillance.

46:10

There are certainly areas of

46:12

overlap. So if you

46:14

look at the recent rise of

46:16

the so-called self-care movement within the

46:18

wellness industry, this idea that you

46:20

should care for yourself and you

46:23

should look after yourself, which comes

46:25

with good intentions. Go to the gym,

46:27

eat healthy, don't overeat, sleep

46:30

enough, take a power nap. We even have

46:32

nap pots in the office maybe and walk

46:34

your 10,000 steps and eat your five portions

46:36

of fruits and veg a day. All

46:39

that is good, but if the real intent

46:41

is for you to be really, really productive

46:43

or to want to stick around work and

46:46

be at the office a lot or to

46:48

return these seemingly well-meaning

46:51

recommendations with your hard work

46:53

and loyalty, then it's normal

46:55

that we are a little bit cynical when we

46:57

hear this advice. And also, if

46:59

our solution to the wellness

47:02

issue and the well-being problems that we

47:04

have seen in the industrialized worlds for

47:06

the last decade or so is to

47:09

just tell people that they should only

47:11

worry about themselves or worry about themselves

47:13

first before they can help others, that

47:16

actually fosters a very selfish and narcissistic

47:18

mindset. Yeah. So what would

47:20

you propose as a solution to this

47:22

phenomenon in order to rehumanize the workplace?

47:25

Well, I think organizations should understand that

47:27

the more people depend

47:29

on technology to do their work and

47:31

be productive, the more they have to

47:33

kind of create cultures that actually provide

47:36

an antidote to that and compensate for

47:38

that. For example, by stimulating analog

47:40

or 3D physical encounters

47:42

between people, by separating

47:45

between activities that might not lead

47:47

to productivity but actually enhance bonding

47:49

and fueling or lubricating the social

47:51

ties that people want with their

47:54

colleagues, irrespective of whether it

47:56

actually makes them more productive or boosts revenues,

47:58

productivities, and profits. think that

48:00

when it comes to caring, you know,

48:03

we have to remember that one

48:05

of the best and I think

48:07

most pro-social ways we have to

48:10

enhance our own happiness and our own subjective

48:12

well-being is to actually be nice and be

48:14

kind towards others, right? So the less you

48:17

think about your own problems and the more

48:19

you try to solve other people's problems, the

48:21

more your problems go away. But

48:23

if I'm listening to this and I run a department and

48:25

I have a, you know, bottom line that I'm expected to

48:28

meet, are some of those things

48:30

in conflict with my ability to meet my productivity goals?

48:33

Well, there is a tension, right? So I

48:35

think generally speaking, it

48:37

is true that on average, other

48:39

things being equal, the

48:42

more engaged and satisfied and happy

48:44

your team is, and of course,

48:46

the more physically fit and energized

48:48

or energetic they are, the

48:50

more productive they will be. But at the

48:52

same time, that overlap is less than 10%. It's

48:55

a correlation of 0.3, which indicates

48:57

a 9% overlap, which means that you'll have

49:00

a lot of people who are extremely healthy,

49:02

their well-being is great and they're very engaged

49:04

but actually they don't add value in terms

49:06

of productivity. And also that

49:08

some of your most valuable high-performing

49:11

or high-potential employees are going

49:13

to be quite grumpy, quite dissatisfied and

49:15

maybe have poor work-life balance

49:17

and, you know, struggle in other areas of

49:19

life. I mean, let's face it, historically, there

49:22

was a tension between people who devote a

49:24

lot of their energies, focus and skills

49:27

on their careers and because of that,

49:29

neglect other areas of social or personal

49:31

life. So I think you have to

49:34

allow for both things and ultimately worry

49:36

less about short-term results and more about

49:38

the kind of culture and climate you

49:40

create in your organization because ultimately people

49:43

are always going to fluctuate. They're going

49:45

to have good years and bad years

49:47

but it's the long-term commitment to a

49:50

strategic goal that actually gives you the

49:52

results in the long-term. People

52:00

gravitate towards and they have mass

52:02

organic adoption because actually they take

52:04

care of boring tasks like proofreading

52:07

texts or emailing colleagues or even

52:09

attending meetings that you don't want

52:11

to attend. The people who

52:13

actually opt in to these tools because

52:15

they see that they can basically stop

52:17

doing things that they don't want to

52:19

do aren't automatically motivated to then reinvest

52:21

the time they save on

52:24

new learning experiences or new kind

52:26

of difficult, effortful ideas. I

52:29

think efficiency is a double-edged sword and it

52:31

can be wonderful but at the same time,

52:34

if at some point we no longer need to think

52:37

and we create something that is like a

52:40

microwave for ideas and we stop actually thinking,

52:42

our brains don't think anymore, then we have

52:44

to wonder what the long-term effects might be.

52:47

Just finally, as unpredictable as the future

52:50

is maybe, what would you say are

52:52

the job skills of the future in

52:54

this automated context? Well,

52:56

I don't pretend to have data on

52:58

the future. I'm always a little bit

53:00

perplexed when I see all these very,

53:03

very granular and detailed calculations of what

53:05

will happen to skills or jobs, etc.

53:09

I think that we need to

53:11

be agile to adjust and adapt

53:13

to whatever comes but it seems

53:15

to me that a reasonable expectation

53:17

is that AI will win

53:20

the IQ battle if it

53:22

hasn't won it already. It will always know

53:24

more about things than we do,

53:26

especially if you count the large number of things

53:29

that can be known even if it doesn't understand

53:31

it. When it comes

53:33

to things like empathy, consideration,

53:35

kindness, self-awareness, people skills, emotional

53:38

intelligence rather than intellectual ability, I think

53:40

we have a real chance to still

53:42

compete and to add value. If you

53:45

think about the manager of

53:47

the past versus the manager of the

53:49

future, in the past they were appointed

53:51

into a management position based on what

53:53

they knew, their qualifications, their hard skills,

53:55

their university credentials. In the future, it's

53:57

probably going to be their ability to

53:59

inspire, to connect with, others to understand

54:01

people and to really give them that

54:03

sense of validation and kindness

54:05

and attention that they will crave, especially if

54:07

they can't even tell whether they're interacting with

54:10

a human or deepfake in some other areas

54:12

of life. Tomás, thanks so much

54:14

for your insights on this. It's been a

54:16

real pleasure. Anytime. Tomás Chamorro-Premusic

54:18

is an organizational psychologist. He's also

54:21

the author of I Human AI

54:23

Automation and the quest to reclaim

54:25

what makes us unique. You've

54:34

been listening to Spark. The show is

54:37

made by Michelle Parisi, Samurit Yohannes, Megan

54:39

Carty, and me, Nora Young. And

54:41

by Joanne McNeil, Alison Pugh, and

54:43

Tomás Chamorro-Premusic. I'm

54:46

Nora Young. You can check out back issues of

54:48

Spark, so find and follow us wherever you get

54:50

your podcasts. For

54:54

more CBC podcasts,

54:57

go to cbc.ca/podcasts.