0:00

At the end of the

0:02

day, that is the

0:04

best route for any

0:07

worker who really worries

0:09

about technology coming and

0:12

being disruptive in their

0:14

workplace. It's just a

0:16

matter of power, right?

0:18

And we've got to

0:21

utilize our powers workers

0:23

together to take this

0:25

on. to Fast Company's The New

0:27

Way We Work, where we take listeners

0:29

on a journey through the changing landscape

0:32

of our work lives and explain exactly

0:34

what we need to build the future

0:36

we want. I'm Fast Company Deputy

0:38

Editor Kathleen Davis. Throughout

0:43

this series on how AI will change

0:45

how we work, we've covered how this

0:47

emerging tech will impact hiring, our daily

0:49

tasks, and how our job performance is

0:52

evaluated. But we haven't yet

0:54

talked about the biggest concern for

0:56

most people when it comes to AI

0:58

and work. Are robots going to take our

1:01

jobs? And honestly, we're right to

1:03

be concerned. According to McKinsey and

1:05

Company, 45 million jobs or a

1:07

quarter of the workforce could be

1:09

lost to automation by 2030 by

1:11

2030. That's just five years from

1:14

now. Of course, the promises that

1:16

AI will also create jobs,

1:18

and we've already started to

1:20

see emerging roles like prompt

1:22

engineers and AI ethicists crop up.

1:25

But many of us have a

1:27

lot of concerns about how AI

1:29

is being incorporated into our fields.

1:31

Should a bot host a podcast,

1:33

write an article, replace an actor?

1:35

Could AI be a therapist, a tutor,

1:37

build a car? Three out of four employees

1:40

say their organization is not collaborating

1:42

on AI regulation and the same

1:44

share say that their company has

1:46

yet to share guidelines on responsible

1:48

AI use. To help us understand

1:50

what protections exist and how employees

1:52

can fight back to ensure their

1:55

jobs are safe is Lorena Gonzales.

1:57

She's the president of the

1:59

California. California, Federation of Labor Unions,

2:02

a former California Assembly woman, and

2:04

has written AI transparency legislation, including

2:06

a law that prevented algorithms from

2:08

denying workers break time. Well, thank

2:10

you so much for being here.

2:12

Well, thank you for having me.

2:14

Can you talk about and explain

2:16

how your job leading labor unions

2:18

in California interacts with national AI

2:20

regulations and trends? So in California,

2:22

I lead the ALCIO, and we

2:24

don't do... directly with national issues,

2:26

the national EFL CIO does. But

2:28

we know that in California we

2:30

have a special responsibility, largely because

2:32

Silicon Valley is here, and so

2:34

much of these tech changes have

2:36

come out of California. We also

2:38

are very aware that little has

2:40

happened or will happen at the

2:42

federal government as far as regulations

2:44

and guidelines. So if we want

2:46

to really set the pace so

2:48

often. within labor laws in general.

2:50

California will pass a law and

2:52

then it becomes more broadly accepted

2:54

first in other what we call

2:56

trifecta states that have pro-labor majorities

2:58

in the House and the Senate

3:00

and in the governor's office. But

3:02

often we'll get picked up as

3:04

well at the national level. So

3:06

although we can't directly or we

3:08

don't really imagine will directly affect.

3:10

federal regulations. We want to make

3:12

sure that we have strong guidelines

3:14

and regulations in California that other

3:16

people can then copy. I'm so

3:18

glad you mentioned that about California

3:20

because it is such unique. place

3:22

not only because Silicon Valley is

3:24

there and that's where a lot

3:26

of this tech is coming from

3:28

but also because the the labor

3:30

world in California is huge you

3:32

know I don't have to tell

3:34

you most populous you know state

3:36

and such you know a major

3:38

union presence in such diverse fields

3:40

you know teachers screenwriters nurses kind

3:42

of everything you can think of

3:44

are there common themes that you're

3:46

hearing from all these kind of

3:48

different groups that you work with?

3:50

There are, you know, when we

3:52

started to first look at regulations

3:54

or guidelines or bills that we

3:56

could. put forward, we thought the

3:58

best way to do it would

4:00

be industry specific. Because when you

4:02

look at AI in the workplace,

4:04

there are very different challenges that

4:06

come in, whether you're talking about

4:08

a health care setting, whether you're

4:10

talking about somebody's like creative work,

4:12

whether you're talking about somebody who

4:15

drives a truck, a big rig

4:17

truck is what I meant there.

4:19

So these are all different issues

4:21

that, especially when it comes to

4:23

safety and privacy. But we realized

4:25

there's some common threads across all

4:27

workplaces as well. One is the

4:29

notion of just robot bosses. We

4:31

did do the First of the

4:33

Nation bill probably about three years

4:35

ago now on algorithmic management. It

4:37

applies only to warehouses and that

4:39

was in response to what we

4:41

saw going on in Amazon and

4:43

to a lesser degree Walmart warehouses.

4:45

And we basically wanted to give

4:47

workers the power. to question the

4:49

algorithm that was speeding up their

4:51

quota, basically their requirement to go

4:53

faster and faster. It also lined

4:55

up with injuries on the job.

4:57

So we thought, this doesn't make

4:59

a lot of sense. You don't

5:01

even have a human interaction. We're

5:03

expanding that this year with a

5:05

bill that would basically say for

5:07

a lot of those interactions, you

5:09

need a human boss. What we

5:11

started with in the warehouse bill

5:13

were really seeing expand throughout different

5:15

types of work. When you're dealing

5:17

with an algorithm, even the basic...

5:19

experience of having to leave your

5:21

desk conversation with that computer. So

5:23

this idea also, what if you

5:25

have to leave work all of

5:27

a sudden because you got a

5:29

call from your kid's school and

5:31

your kid is very sick or

5:33

there's some emergency. Taking away the

5:35

human element obviously has a structural

5:37

problem for workers, but it has

5:39

a humanity problem as well. We

5:41

also know that with algorithm management

5:43

that the computer is making certain

5:45

decisions and relying on certain assumptions

5:47

that we may not even. allow

5:49

openly managers to rely on, right?

5:51

Your race, your gender, some assumptions

5:53

that the computer has been programmed

5:55

to rely upon, and there's no

5:57

regulation, there's no liability, there's no

5:59

ability, there's no ability to even

6:01

know what's going into the computer

6:04

to know what the outcome and

6:06

what the computer is trying to

6:08

do. So having a computer is

6:10

a boss. A robot is a

6:12

boss, if you will, is one

6:14

of the ways we thought in

6:16

every setting, everybody kind of identified

6:18

that as problematic, dealing with the

6:20

idea of algorithmic management. So that's

6:22

one bill we're going to do.

6:24

Privacy is another. I mean, just

6:26

straight out privacy. And we started

6:28

seeing what a lot of companies

6:30

were putting out into the ether

6:32

about what corporations could utilize. So

6:34

we think there are certain spots

6:36

in work, no matter where you

6:38

work, that should be private to

6:40

an individual. That's going to be

6:42

a bathroom. It's going to be

6:44

a break room. In your car

6:46

and in your home, if those

6:48

places aren't places of work, you

6:50

should have a basic sense of

6:52

privacy. And what we used to

6:54

believe is privacy, like, well, they

6:56

can't, they can't watch me. They

6:58

can't listen to me. These are,

7:00

you know, it allowed me to

7:02

have some privacy. No, we've got

7:04

to go much further than that.

7:06

It has to be heat mapping.

7:08

It has to be all the

7:10

tools that are being marketed to

7:12

these corporations. We need protections against.

7:14

Your employer should not be able

7:16

to know every person you're talking

7:18

to in the restroom because you're

7:20

both wearing some device that tells

7:23

them that you're talking to each

7:25

other. A boss definitely should not

7:27

know who you're talking to in

7:29

a break room. Of course, as

7:31

a union organizer, we really think

7:33

that's an important feature of work

7:35

is having the right to. When

7:37

you're off the clock, if you

7:39

will, to interact with your coworkers

7:41

in a way that may result

7:43

in unionization, and when they're able

7:45

to use surveillance mechanisms to stop

7:47

that, that's dangerous. So we have

7:49

a bill on that as well.

7:51

And the third area that we

7:53

think really applies to everybody is

7:55

just a basic data issue. And

7:57

we haven't released that bill yet.

7:59

We're working on the language of

8:01

that bill. But you know, the

8:03

data that is being taken from

8:05

workers without their knowledge, their personal

8:07

data, their facial features, their whatever

8:09

it is, the right to understand

8:11

what is being taken by a

8:13

computer or by AI as you're

8:15

doing the work, sometimes to replace

8:17

you, sometimes to evaluate you. That

8:19

data goes into forming something else.

8:21

Now this became an issue really

8:23

obviously during the rider strike. It's

8:25

so much clear and obvious to

8:27

people when you're saying you're taking

8:29

their work, feeding it into computers,

8:31

it's coming out somewhere else. But

8:33

this comes in different forms. We've

8:35

heard it from longshoremen who say

8:37

the computer works side by side

8:39

to try to mimic the responses

8:42

that the worker is giving. It

8:44

might be to tide levels or

8:46

to dangerous situations. The workers should

8:48

have the right to know that

8:50

they're being monitored, that their data

8:52

is being taken, and there should

8:54

be some liability involved. So those

8:56

are kind of the broader issues

8:58

we've identified in ones we're going

9:00

to tackle with legislation this year

9:02

or try to tackle. Thank you

9:04

for sharing those. Those are great,

9:06

and that is a... a great

9:08

kind of overview of the way

9:10

that AI can, similar AI can

9:12

impact, you know, different industries. I'd

9:14

love, I'd love to hear more

9:16

from you about kind of more

9:18

specific AI in industry since we've

9:20

seen such as you mentioned, we've

9:22

seen such high profile labor. disputes

9:24

in the last year regarding AI.

9:26

Obviously the screen actors was a

9:28

big one. The auto workers had

9:30

an AI provision in their labor

9:32

dispute. Is there a way to

9:34

regulate AI implications within industries or

9:36

is that kind of a case-by-case

9:38

labor issue and these these larger

9:40

bills are kind of the only

9:42

way to tackle it in a

9:44

larger way? There's absolutely ways to

9:46

tackle it industry by industry and

9:48

we've done that some successfully, some

9:50

less successfully. So we got some

9:52

provisions for SAG After and for

9:54

our actors in legislation last year.

9:56

That was a positive. One of

9:59

the things we've done in the

10:01

past and we got a bill

10:03

through and signed was basically when

10:05

any new technology is deployed in

10:07

public transit that it becomes a

10:09

a mandatory subject of bargaining. So

10:11

that's good where you have a

10:13

union, where the union in a

10:15

workplace, whatever the technology is, can

10:17

have a discussion and can be

10:19

forced, management's force to discuss it

10:21

with workers. That's less protective, but

10:23

it's important. We've run a couple

10:25

of bills where we really are

10:27

talking about human operators, and this

10:29

is something that we have not

10:31

been able, we've gone it through

10:33

the legislature with broad bipartisan support,

10:35

but not have, haven't been able

10:37

to get the governor to sign.

10:39

And that's like, a big rig

10:41

or a school bus or a

10:43

big barge should not be operating

10:45

just simply with AI without a

10:47

human operator. So we want to

10:49

do that, number one, to save

10:51

jobs, but number two for safety

10:53

reasons as well. And I think

10:55

most people don't want to be

10:57

on a highway with a big

10:59

rig that has no human operator

11:01

as a backup. We have some

11:03

pending issues, I think, going on

11:05

with pilots, right, where you have

11:07

airlines that want to go to

11:09

one pilot. and a computer rather

11:11

than the two pilots that we

11:13

now have and rely upon to

11:15

make sure that we're safe. These

11:18

are anti-consumer, they're anti-worker, and they're

11:20

things that we think can be

11:22

and should be regulated in legislation.

11:24

There are a lot of things

11:26

in health care that we've been

11:28

looking at in trying to figure

11:30

out how to get at. Health

11:32

care is a special industry that

11:34

AI has been very helpful. There

11:36

are things that we want AI

11:38

to do. Of course, I'm a

11:40

cancer survivor. I would like the

11:42

computer to view all of my

11:44

images along with my doctor to

11:46

see if they can pick something

11:48

up that my doctor couldn't. Of

11:50

course that's a great thing, but

11:52

we also have seen AI and

11:54

computers go a little rogue in

11:56

the hospital sitting where they're insisting

11:58

on certain tests or certain procedures

12:00

that the nurse knows is inappropriate

12:02

and the doctor doesn't have to

12:04

okay. The nurse gets into a

12:06

position where it's easier, in fact

12:08

it's much harder to challenge the

12:10

decision of the computer than to

12:12

simply over test folks. That's that

12:14

that can be dangerous as well.

12:16

We've also been looking at who

12:18

actually is providing the information. So

12:20

in California, for example, you need

12:22

a medical license to practice medicine,

12:24

but the computer doesn't have a

12:26

medical license, and we don't even

12:28

know who's programming the computer and

12:30

who has a medical license and

12:32

what the person behind the information

12:34

being given, how credible it is.

12:37

And so these are all questions.

12:39

And last but not least, of

12:41

course, liability. So if you go

12:43

to the doctor and there's some

12:45

really bad screw-up, of course, there's

12:47

medical malpractice and you have the

12:49

right to sue, you have the

12:51

right to some kind of recourse,

12:53

in these situations where the computers

12:55

are making the decision and doing

12:57

the work, who's liable? I mean,

12:59

that's even something that hasn't really

13:01

been established, and that's something that

13:03

continues to be a question. Those

13:05

are all great points. It's also,

13:07

you know, it's really the consumer,

13:09

the customer, the patient, the worker,

13:11

and as you bring up in

13:13

a liability issue, you know, the

13:15

company itself. And that's something that

13:17

we haven't touched on yet. You

13:19

know, we've been talking a lot

13:21

about legislation and regulation. I opened

13:23

this episode with some stats about

13:25

how the majority of employees say

13:27

that their companies are not collaborating

13:29

on AI regulations. or that their

13:31

companies haven't even kind of shared

13:33

guidelines for responsible AI use. If

13:35

I'm a business leader and I'm

13:37

listening to this, what's your guidance?

13:39

Well, if you have a union

13:41

in your workplace, absolutely, you should

13:43

be getting together me and conferring

13:45

over any new technology. A lot

13:47

of times that's in a contract,

13:49

but even if it's not, it's

13:51

just good practice. But if you're

13:53

an employer without a union, that

13:56

should be a discussion that you're

13:58

having with your employees in an

14:00

open discussion, not a discussion whereby

14:02

you've already made the decision, you're

14:04

going to go in a tech

14:06

route in an AI route without

14:08

actually talking to the people doing

14:10

the work and seeing how that

14:12

will affect their... job, their livelihood,

14:14

and what they're doing. I think

14:16

so often, especially those at the

14:18

very top, get sold on new

14:20

technology as being cool and being,

14:22

you know, innovative and being able

14:24

to do things faster and quicker

14:26

and not really going through the

14:28

entirety of what these jobs are

14:30

and not really imagining what on

14:32

a day-to-day basis that worker has

14:34

to deal with. And so I

14:36

think that is really important is

14:38

to sit down and actually talk

14:40

to the workers and empower the

14:42

work. to have a say in

14:44

the technology. Yeah, I think that's

14:46

true. And I, you know, in

14:48

the first episode of this season,

14:50

I talked to an AI ethicist

14:52

who was talking exactly about that.

14:54

I think the way that AI

14:56

technology is sold is often, you

14:58

know, not only is this an

15:00

amazing development, but is in these

15:02

very technical terms of people like,

15:04

well, I don't understand it, but

15:06

I'm sure it's good. Yeah. And

15:08

kind of, you know, having that

15:10

really human skill, right of critical

15:13

thinking. to ask. Absolutely and I

15:15

think in public employment we're really

15:17

concerned right because so often what

15:19

will happen is these city managers

15:21

or our county clerks the appointed

15:23

positions or the elected positions go

15:25

to these conferences and you have

15:27

some presentation on how this can

15:29

help you reduce the workforce or

15:31

save money and they buy into

15:33

it. But they buy into it

15:35

and then reduce a workforce that

15:37

actually does that critical work. And

15:39

at the end of the day,

15:41

if that technology doesn't work, if

15:43

it doesn't work out or you

15:45

don't have the ability it costs

15:47

too much to fix it or

15:49

replace it or there's problems, you've

15:51

then gotten rid of an entire

15:53

workforce that actually can make sure

15:55

those vital services are happening. You

15:57

know, there should be some guardrails

15:59

about how you change, especially when

16:01

you're replacing people in jobs and

16:03

that institutional knowledge of how to

16:05

get a job done. And we

16:07

talk about vital services, whether it's

16:09

ensuring that individuals can sign up

16:11

for their health care, like MediCal

16:13

and California, or get their food

16:15

stamps, or to make sure that,

16:17

you know, the police come when

16:19

you call 911, all of these.

16:21

can't risk losing human beings who

16:23

know how to do this job

16:25

because somebody went to some conference

16:27

and saw a really cool computer

16:29

that seems to be able to

16:32

do it without looking at the

16:34

full cost and the the applicability

16:36

in any situation. Well, in speaking

16:38

of this technology, again, you know,

16:40

you're in California, a lot of

16:42

this technology is coming out of

16:44

Silicon Valley. What are you seeing

16:46

about how the AI industry is

16:48

interacting with the rest of the

16:50

state, with the unions, with the

16:52

lawmakers? You know, I don't see

16:54

a lot of, I went to

16:56

CS in Las Vegas this year,

16:58

and we did have a little

17:00

tour, it was all, I went

17:02

with union leaders, it was fascinating,

17:04

but I don't see that they're

17:06

actually reaching. out. I think they're

17:08

trying to give it a finished

17:10

product and saying this works. But

17:12

I've got to be honest too.

17:14

I've been in government before in

17:16

between I was a labor leader

17:18

and then I was a legislator

17:20

and then back to being a

17:22

labor leader. And we were having

17:24

tech problems in California because our

17:26

computer system is on COBOL. You

17:28

know, it's like, are we going

17:30

to have AI? One of the

17:32

problems is you didn't have anyone

17:34

that could fix an old computer

17:36

system. And so we bought a

17:38

new computer system. with Oracle and

17:40

it didn't talk to the old

17:42

computer system. All right, well. Can

17:44

AI fix that? Because if not

17:46

all you're going to do is

17:48

put one more thing on top

17:51

of it. I know people get

17:53

excited and want the newest, best,

17:55

quickest thing, but the reality is

17:57

our infrastructure is dated. And so

17:59

what is the plan there? I

18:01

don't think there's been a lot

18:03

of interaction. Instead, it's just a

18:05

finished product like here. You could

18:07

use this rather than working with

18:09

procurement and with the unions, the

18:11

workers that actually are doing that

18:13

are work the best to perform

18:15

efficiencies to ensure that there's safety

18:17

and to ensure that there's still

18:19

a human operator. Are there any

18:21

cases, I mean, you mentioned it

18:23

a little bit when your health

18:25

care example, are there cases where

18:27

that you've seen the AI can

18:29

be good for workers? I'm sure

18:31

there are, like I think about

18:33

this a lot, I have to

18:35

say. When we think about who

18:37

has the most amount of technology,

18:39

ability to deploy AI, has the

18:41

resources, we often think of Amazon,

18:43

right? Amazon has these massive warehouses,

18:45

they have invested in a lot

18:47

of technology, they understand it. We

18:49

saw during COVID, when they started

18:51

screening people for their body temperature,

18:53

which is now morphed into screening

18:55

body temperatures to see who's talking

18:57

to each other. So we know

18:59

that there's a lot out there

19:01

that can be done. But what

19:03

amazes me is they haven't used

19:05

all that expertise, all that knowledge,

19:07

to figure out how to reduce

19:10

injuries on the job. So you're

19:12

asking like, why if you have

19:14

the ability to make a job

19:16

safer for the actual worker that

19:18

is still being deployed to do

19:20

it, why are you not? using

19:22

your technology to do that. And

19:24

I think that's a valid question.

19:26

Obviously, you know, look, when you

19:28

look at an airplane being driven,

19:30

we like that the computer system

19:32

is a backup, but we also

19:34

want the human operators on there

19:36

to have eyes on it as

19:38

well. If you look at some

19:40

of the... kind of famous like

19:42

saved crashes it's a human being

19:44

overriding the computer system that saved

19:46

it on a day-to-day basis. I

19:48

think the computer system, the technology,

19:50

can keep, especially in bad weather,

19:52

you from hitting another plane or

19:54

hitting a mountain. Those are good

19:56

things, right? So I think technology

19:58

can absolutely be good, and it

20:00

has been good when it comes

20:02

to directions for drivers, you know,

20:04

understanding when something's closed and open

20:06

so that they're not wasting time

20:08

going down a street that now

20:10

has been blocked off. These are

20:12

all good things. It's just a

20:14

matter of utilizing the good with

20:16

utilizing the agency of the individual

20:18

workers as well and figuring out

20:20

how to mold those two and

20:22

unfortunately we don't have a lot

20:24

of corporations who care deeply about

20:27

about that worker feature. And sometimes

20:29

when we do have corporations that

20:31

care about the safety, it's just

20:33

because they know, for example, nobody's

20:35

gonna get on a plane that

20:37

feels like it's gonna crash. Right.

20:39

It comes down to profits and

20:41

money, and that is the motivation

20:43

of most corporations, and what's wrong

20:45

with allowing them to, you know,

20:47

later make these decisions on technology

20:49

in the workplace. So I asked

20:51

your opinion on what business leaders

20:53

should think, and as I've been

20:55

covering on this season, AI is

20:57

coming for nearly every workplace, whether

20:59

we like it or not. What

21:01

is your advice to the average

21:03

American worker on how to think

21:05

about AI in their working life

21:07

and what they can do if

21:09

they're, you know, maybe justifiably afraid

21:11

about it? Well, my advice is

21:13

always to form a union. You

21:15

know, we have, you can Google

21:17

how to form a union. We

21:19

have a website, unionize california.org. The

21:21

bottom line is an individual worker

21:23

is going to have a hard

21:25

time fighting tech changes or protecting

21:27

their job or really taking on

21:29

the boss. But when workers get

21:31

together and make demands and make

21:33

changes, then real change can happen.

21:35

And we've seen protections in workplaces

21:37

that are unionized that far out.

21:39

legislation that far outpaced any other

21:41

workplace and that's because they can

21:43

collectively bargain. At the end of

21:46

the day that is the best

21:48

route for any worker who really

21:50

worries about technology coming in being

21:52

disruptive in their workplace. It's just

21:54

a matter of power right and

21:56

we've got to utilize our powers

21:58

workers together to take this on

22:00

because now the boss is doing

22:02

what the boss has always done

22:04

and that's trying to figure out

22:06

how he or she can get

22:08

richer at the so often their

22:10

workforce and now it's in collaboration

22:12

with a computer, but the power

22:14

of individual workers is still always

22:16

going to be bigger. Well, I

22:18

think that's kind of a perfect

22:20

segue into my last question, which

22:22

is I always like to end

22:24

on a like best and worst

22:26

case scenario. So what are your

22:28

kind of worst case and best

22:30

case scenarios for AI in 2030

22:32

in five years from now? I

22:34

think the worst case would be

22:36

that we don't regulate it. It's

22:38

been interesting to kind of watch

22:40

that these tech companies really come

22:42

in and number one tells we

22:44

don't understand it so we can't

22:46

regulate it. We should study it.

22:48

All of this so that they

22:50

can continue to grow and be

22:52

too big to fail, too big

22:54

to take on. And so I

22:56

think if we don't do our

22:58

job and actually regulate the use

23:00

of technology in the next five

23:02

years, then they may be too

23:05

big to take on. And the

23:07

computer can be more powerful than

23:09

than actual human beings. That would

23:11

be very dangerous. I think what

23:13

What we don't want to do

23:15

is to be scared into paralysis

23:17

and not do anything about it.

23:19

And that's what we're trying to

23:21

do in California is say, how

23:23

do we empower workers both in

23:25

their workplace and through the legislature

23:27

to say this isn't, we don't

23:29

have to accept this. This, you

23:31

know, we can regulate tech and

23:33

we should. And you don't have

23:35

to know how to code or

23:37

understand how AI is created to

23:39

regulate it. You can say that

23:41

there are guardrails and there are

23:43

safety in their regulations and not

23:45

understand how. AI comes about. That

23:47

has been smoke and mirrors and

23:49

so I think that it would

23:51

be the worst case scenario that

23:53

we're just paralyzed that nothing happens

23:55

that there are no regulations that

23:57

there's no good bargaining that goes

23:59

on in workplaces. I know that

24:01

won't happen because we have good

24:03

unions that are working on this

24:05

now. The best case scenario is

24:07

that we actually regulate this, that

24:09

we set up guardrails that make

24:11

sense, that we protect and preserve

24:13

human jobs. That is a noble

24:15

thing. I mean, I don't think

24:17

anybody wants a jobless society or

24:19

to have just certain jobs that

24:21

it makes economic sense only to

24:24

have, right? We do the value

24:26

of humans and humanity, the value

24:28

of having a human operator. I

24:30

think we can. introduce the best

24:32

of technology into the workplace with

24:34

a good collective bargaining agreement, with

24:36

laws in place that regulate it,

24:38

that have guardrails, and then with

24:40

institutional collective bargaining in places where

24:42

their unions to ensure that that's

24:44

all deployed correctly. That would be

24:46

the best case scenario. Well, I

24:48

can't think of a better way

24:50

to end it on a best

24:52

case scenario. Thank you. This is

24:54

the last episode in our mini-series

24:56

on how AI is changing our

24:58

jobs, but we are by no

25:00

means done with this topic. As

25:02

we've covered, AI is here to

25:04

stay and we are just starting

25:06

to understand how it can be

25:08

both used and misused in all

25:10

areas of our lives. For me,

25:12

an AI skeptic, my biggest takeaway

25:14

throughout this series has been the

25:16

need for human intervention. We know

25:18

that AI is only as good

25:20

as the data it's trained on,

25:22

but its implementation is also only

25:24

as effective as the humans using

25:26

it. For hiring, this means using

25:28

AI to write interview questions, but

25:30

then editing and rewriting them yourself,

25:32

or making sure that the keywords

25:34

that AI is screening for don't

25:36

leave. out those with

25:38

transferable skills. skills.

25:40

For implementing AI

25:43

into daily work,

25:45

it means vetting

25:47

systems to make

25:49

sure they actually

25:51

live up to

25:53

their promises. live up

25:55

For performance reviews,

25:57

it means using

25:59

AI to analyze

26:01

data, but not

26:03

replace human conversations not

26:05

replace The bottom

26:07

line is I

26:09

see it, AI

26:11

right now feels

26:13

less like a

26:15

total game feels less

26:17

like a new

26:19

tool that we're

26:21

just learning how

26:23

to use. tool that we're

26:25

just learning how to use. Well,

26:30

this season is done. We season is done, we

26:32

will be back with new episodes

26:34

soon featuring some of your most pressing

26:36

workplace questions. Be sure to subscribe

26:38

to The New Way We Work wherever

26:40

you listen so you never miss

26:43

an episode. never if you liked this

26:45

episode, leave us a rating or review

26:47

on Apple leave us a rating review Way We

26:49

Work is hosted by me, Way Davis,

26:51

and produced by me, Kathleen Cody Nelson

26:53

and Joshua by Henry Chardonnay, Torres. and

26:56

Joshua Christensen with mixing

26:58

by Nicholas Torres.