0:03

You're listening to a CNA podcast.

0:10

Hi, welcome back to the Work It podcast

0:12

with Tiffany and Gerald. Now, not too

0:14

long ago, someone asked me if

0:16

my job can be easily replaced by artificial

0:19

intelligence. If you asked me that

0:21

question a year ago, I would say no

0:23

because I didn't think AI can replicate

0:26

my voice that perfectly or

0:28

write a punchier introduction to

0:30

the start of this episode. But

0:32

these days, um I'm not so sure about that

0:34

because I think AI might soon be able to

0:36

write more creatively than I can.

0:39

But Gerald, do you use AI

0:41

in your line of work? Yes, lots

0:43

of it. In the process of guiding

0:45

our clients with resume writing

0:47

and job search processes, AI

0:49

plays a really, really big part. It has made a

0:51

lot of the painful parts of the process a

0:54

lot more easy to live with. But I do also

0:56

know that there are people who come up to me and also

0:58

ask like, OK, now if I'm looking for a job,

1:01

and then I'm looking at all these jobs that have AI

1:03

requirements, right? So what do I need? How

1:05

do I keep up to date? With all these AI

1:07

changes. Yeah, because I think for the average

1:09

person, it looks like if you know how to

1:11

use chat GPT just to shave

1:14

off some time from writing an

1:16

email or writing maybe a

1:18

proposal, a deck, that

1:20

should be enough, right? Yeah, I think that's

1:22

the most obvious uses of

1:24

AI, but I think today we want to go a

1:26

bit deeper to understand how deep

1:29

really AI can penetrate into different job roles

1:31

and how we can keep up with that changes.

1:33

So today I'm Happy we have with us in studio,

1:35

Ku Sing Ming, head of Learn AI

1:37

at AI Singapore, to discuss more. Welcome

1:40

Sing Ming. Hello, hi, thanks for inviting

1:42

me. Singing, we've often heard that

1:44

AI is reshaping our

1:47

job roles and that it's not exactly

1:49

going to be a niche thing if you have AI

1:51

skills. In fact, it's going to be considered a core

1:53

competency for many people,

1:55

many jobs. So what exactly

1:58

are these AI skills that

2:00

companies want their work? You're

2:02

right that more and more companies

2:04

are looking for their staff

2:07

or future hires to have some

2:09

form of AI skills, but then

2:11

in itself, the AI skills can

2:13

be a spectrum. It will largely

2:16

depend on the type of roles, but

2:19

I will kind of break it down into a

2:21

non-technical requirement and a technical

2:24

requirement. So for non-technical

2:26

requirement, it focus less

2:28

on Your ability to code,

2:31

your ability to actually create AI algorithm,

2:34

but rather your ability to use

2:37

AI tools that are quite

2:39

relevant and useful to The

2:42

particular company's business area

2:44

or the type of work and job

2:46

functions that the company is looking for.

2:49

For example, looking at tools

2:51

to streamline workflows, to

2:53

get things done better, faster,

2:56

to even enhance creativity

2:59

also. To your point, to talk about

3:01

the opening intro itself, you

3:03

definitely would have created own

3:06

competency in creating punchy

3:08

taglines, intros all that. But

3:11

it will be very useful for you to have

3:13

an AI tool to help you to write up the

3:15

draft and then you look

3:17

at the draft and layering with your

3:20

own tonality, your own personality

3:22

that is very much your own. So AI

3:24

will be seen more as a digital

3:27

assistant. There are so many

3:29

tools out there and many of them are actually

3:31

free. So can the company

3:33

leverage on this type of resources

3:35

out there? And they will be looking at

3:37

people who are very effective

3:40

and very creative in using those tools to

3:42

achieve what the company wants and what the job

3:44

requires. So that's the non-technical

3:47

side. Then the technical side would be for

3:49

the technical side itself, it is

3:51

becoming more specialized.

3:53

The requirements are also deeper. There

3:56

is still requirement for certain

3:58

types of companies, especially digital companies,

4:01

to create a gorithm. That

4:04

will help to power their AI

4:06

solution or digital solution, or

4:08

they may have existing digital

4:10

platform or digital solutions that will benefit

4:13

very well from having autonomous

4:15

AI agent operating inside. So

4:18

companies will still be looking for people

4:20

with coding proficiencies, for example,

4:22

like in Python, they definitely

4:25

is looking for people with

4:27

data engineering skills. Because

4:30

today, how do you manage all the data

4:32

sets that comes in and also

4:34

the ability for you to, after you

4:37

create an AI solution, how

4:39

do you integrate it into your business

4:41

back end? The term machine learning

4:43

operations, the discipline of

4:45

operating an AI model in

4:48

a business environment. Production

4:50

environment. I think those would be

4:52

very important requirement that

4:54

companies will look for when they have

4:56

technical requirements for such people.

4:59

I think that's a very helpful breakdown. I see as

5:01

like AI users and AI builders,

5:03

the users are the ones who are using

5:06

the benefit of the platform to augment their

5:08

work to make things better, faster. And then

5:10

you've got the builders that they need to have deeper expertise

5:12

to build the technology, build the algorithms,

5:14

and then weave it into the business processes itself.

5:17

So Sing Ming, for the majority of the industries

5:19

out there, would you see that the demand

5:21

is more for AI builders or

5:24

AI users? I like

5:26

the way you're describing it. Maybe I'll just

5:28

contribute to a point. Singapore, we just

5:30

released our new national AI strategy,

5:33

version 2.0. We call it NAIS 2.0

5:36

for short. We actually identify

5:39

3 types of AI talent archetype

5:41

in Singapore. So, we

5:43

have our AI users and

5:46

what you call builders are actually AI practitioners,

5:48

practitioners, right? And then the

5:50

highest is AI creators. So 3 types.

5:53

AI users, AI practitioners, and

5:55

AI creators. And this

5:57

is a wonderful framework that not

5:59

only guide a person, whether

6:01

he or she wants to belong

6:04

to each category, it also

6:06

helps for organization. To

6:09

think about how many AI users do

6:11

they want, how many AI practitioners

6:14

should be part of the organization, and

6:16

do I need the top AI creators

6:19

also in my companies to help me to

6:21

create new ways of solution,

6:23

products and services. Does that make me more competitive.

6:26

Do you think most big companies need to have

6:28

AI creators right at the top,

6:31

at least giving the company a form of strategy

6:33

to take the company forward? I believe so.

6:36

You need to optimize it for your particular

6:38

industry. For example, If

6:40

you look at a digital first

6:43

industry, which means companies

6:45

that operate on digital platform

6:48

and a lot of their products and services

6:50

depend on customer inputs

6:52

or customer visiting their platform,

6:54

for example, like Lazada, e-commerce

6:57

platform itself. Then these

6:59

companies would definitely want to have

7:01

more AI creators to create

7:03

new and novel way of delivering their

7:05

services. They would probably

7:07

also want to have a lot of AI

7:09

practitioners. So the ratio

7:12

of AI creators, AI practitioners,

7:14

and AI users could

7:16

be evenly distributed across

7:18

digital first industry. Now,

7:20

on the other hand, let's say Yakunon, Yaoon

7:23

is in the business of serving

7:26

quality breakfast. Do they need AI

7:28

creators? Do they need they need AI

7:30

builders maybe? Do they need AI

7:33

users themselves? Yeah, could be as

7:35

well, right? So then the ratio will be different,

7:37

right? You definitely want to

7:39

have your staff intelligently thinking

7:42

about how can I do my job better? Was

7:44

there an AI tool somewhere based on interaction

7:46

with the customers, you can feedback to the company.

7:49

Can we have some kind of smart tools powered

7:51

by AI to do certain things? So AI users

7:53

will be one. Practitioners then

7:55

would be probably the company

7:57

having sort of a little development

8:00

team or someone will actually know enough

8:02

about technical details to work

8:04

with either AI startups or

8:06

AI solution providers. AI

8:09

creators, to your point, maybe there isn't

8:11

a need. They can just buy the model

8:13

from somewhere else, right? You're right.

8:15

So it's either which ROI

8:17

serves you better. So you perfectly

8:20

encapsulate that ratio. Do you build

8:22

or buy? Yeah, that's right. I think when

8:24

we look at the AI practitioner,

8:27

let's say at a practitioner level, in

8:29

the different industries, what kind of job roles

8:31

can we expect the practitioner to create?

8:34

Let me give us some examples of industries or

8:36

businesses so that we can visualize better. Some

8:38

examples, it could be HR

8:40

could be using the chat GPT.

8:43

Generative AI tools itself

8:45

to look at generating

8:47

certain types of JD that

8:49

could be more targeted, that could be more

8:52

specific. So, you could be generating

8:54

more customized and more tailored one

8:56

to screen out better applicants. Logistic

8:59

companies are constantly interfacing

9:02

with their customers, could be

9:04

angry customers or the suppliers

9:06

or that. Could there be a form of automated

9:09

email replies because several

9:11

of the email transactions could be just updating

9:13

of information or simple queries

9:16

or that. So that could be one, which

9:18

then also lend into customer facing

9:20

functions across a variety of industries

9:23

itself. This could be some of the

9:25

broader areas which

9:27

a lot of interaction can be

9:29

automated away. If it's

9:31

more of simple reply, simple clarification.

9:34

On a more deeper aspect

9:37

of application and usage, right,

9:39

you could be using GAI

9:41

tools and more advanced

9:43

AI algorithms to assist in your

9:45

product planning. You could also

9:47

be using it to be more

9:49

creative. So it's

9:52

not about AI taking creative

9:54

away from the industries,

9:56

but rather enhancing creativity.

9:58

So we can have companies

10:00

that help in terms of looking

10:03

at defect reduction. So

10:05

we have a multinationals that was

10:07

looking at manufacturing. So

10:09

for them, manufacturing production you,

10:12

defect you is very important. Can

10:14

you build a machine learning algorithm

10:16

or predictive system? To help

10:18

to predict which type of product lines

10:21

would have a higher defect yield, finding

10:23

out the cause, what causes

10:25

that. So when you find out the cost

10:27

and you can lower the defect rate,

10:29

you can actually apply that to other product

10:31

lines, makes for better product planning.

10:34

So, so even in the non-technical

10:36

side, product planning itself, companies

10:38

that are looking at future consumer demands

10:41

or that could use AI

10:43

tools to vastly integrate

10:45

huge amount of information from

10:47

the internet and consolidate

10:50

them. And provide either certain

10:52

trend or analysis or even

10:54

suggestion itself. For the creative

10:56

industry itself, I came across this company

10:58

called Mighty Bear. So Mighty Bear is

11:00

a games company. They

11:02

also help to look at providing some

11:05

creative design for clients or that.

11:07

And when they talk about generative

11:09

AI Dale stable diffusion,

11:12

all that, it is the opposite

11:14

of, oh. Clients now will

11:16

be using Dale or stable diffusion

11:18

and they don't want to engage us. In fact, it's actually

11:20

the opposite because

11:22

by being able to use

11:24

those tools, they are actually

11:27

generating more designs for

11:29

the client to choose from. So

11:31

in effect, the opposite has happened.

11:34

They have become more creative, they

11:36

have become more responsive

11:39

to clients' needs. And all

11:41

in all, it actually enhance the client's

11:44

experience itself. So if somebody

11:46

says, OK, you know what, I've tried AI

11:48

and I'm really not as good because

11:50

this whole computer thing doesn't quite jive

11:52

with me or I find it very hard to

11:54

learn. So then what do you say to somebody

11:57

like this who feels that they may

11:59

not be able to pick up AI skills as

12:01

quickly as the people around them.

12:03

Hm. The term AI

12:06

skills can be quite broad.

12:08

Many will interpret it as, oh,

12:11

I need to learn coding, I need to

12:13

understand what is machine learning, I need to

12:15

understand what's deep learning, structured unstructured

12:18

data, the technicalities of it,

12:20

and For someone, a working

12:23

professional who has spent years in

12:25

their own subject matter, honing

12:27

their own skills itself, it could be quite a

12:29

leap. For the average worker is I

12:31

still love what I do. I'm

12:34

being asked by the government, by the

12:36

company to pick up AI skills. Actually,

12:39

what the company is looking at and

12:41

what the person should be thinking about

12:43

is what kind of AI tools

12:45

can I use to make me become

12:47

better. Yeah, I agree with what Ze Ming

12:49

is saying here. I saw this report recently about

12:52

social workers because a part of social workers,

12:54

they have to do case notes and case notes takes

12:56

a lot of time to record down the details to

12:58

conceptualize about a client's background. They

13:00

found a way to use AI to create the case

13:02

notes. So every conversation is transcribed.

13:05

It comes out in bullet points and it's formed into a

13:07

template for case notes and then wow, with the

13:09

time savings, the social worker can

13:11

spend more time with the families, spend more time talking,

13:14

investigating a little bit more, and helping on the

13:16

ground. So I thought that, wow, this is a

13:18

really good example of how it can

13:20

really augment and allow us to

13:22

personalize a lot more. If

13:25

AI can help to take away the admin

13:27

part of it, then it actually reduces

13:30

in the long run that burn out as well because I

13:32

think a lot of people who are in healthcare or

13:34

in social work, they are also saying that

13:36

we want to be in the field to do the work, but

13:38

sometimes it's like the paperwork that really gets

13:40

to us. And now if we equip

13:43

ourselves with these skills that can make our jobs

13:45

even easier, even more streamlined,

13:48

then we can do the thing that we signed up for the

13:50

things that we really, really love. I think

13:52

we've gone beyond the conversation of whether

13:54

AI can replace our jobs. I think

13:56

here is where we are asking how

13:58

can AI really partner us to do our

14:00

jobs better? How can we at

14:03

least get baseline competent

14:05

in the next 35, 10

14:07

years in our jobs. So thank you so much

14:09

for coming on and sharing with us. Thank you for

14:11

inviting me to speak. Hi,

14:17

we're back with our Ask Me Anything

14:19

segment where we take a work-related question

14:21

that you've said. So let's start.

14:24

Today's question was sent in by Sandra.

14:26

Sandra requested and submitted an

14:29

application to work from home for 2.5

14:31

days under the new flexible

14:33

working arrangement guidelines. Just to recap,

14:36

this new policy kicked in last December

14:38

and employees may submit

14:40

a request. For flexible working

14:43

arrangements, but this is

14:45

still subjected to the employer's approval.

14:48

Now, Sandra is asking for this

14:50

because she says her husband will

14:52

undergo cataract surgery soon and

14:55

she needs to be home during the day

14:57

to care for him while fulfilling her

14:59

work duties. Sandra says

15:01

that her request was rejected and her

15:03

boss told her to apply for the 2.5

15:06

days, but this time of annual

15:08

leave instead. The reason given,

15:10

encourage employee to clear annual

15:12

leave. Yeah, I think

15:15

the challenge here that Sandra faces is really

15:17

what qualifies as leave and what qualifies

15:19

as flexi work. With the new FWA

15:22

kicking in, I'm sure we will have a lot more of these

15:24

sorts of requests because it's becoming

15:26

unclear to people when should we be asking

15:28

for flexible work and when should we be taking leave.

15:31

So I think for me, the difference here is really

15:33

from an employee perspective, the difference. Here are

15:35

3 things. One is our attention to our work.

15:38

I'll be able to dedicate and devote

15:41

our attention fully to our work. And also,

15:43

the second thing is our availability to respond

15:45

to contingencies to request on the

15:47

job. And the third thing, of course, is the

15:49

capacity to complete the work that we said

15:51

we would do. So I think when we think about flexible

15:54

work arrangements, we have to take these 3

15:56

things in mind. So that we know

15:58

whether it's, are we still able to

16:01

perform, to cope, to pay attention to our

16:03

work while we are away, let's

16:05

say in Sandra's case away from the workplace.

16:07

Yeah, so the three factors again just to recap, attention

16:10

to work, availability to

16:12

respond and capacity to complete. So

16:14

these three factors need to be taken into

16:16

account. Yes, because when you are working from home

16:19

Still at work, right? So you'll definitely

16:21

need to be within reach from your team, from your bosses.

16:24

You need to be able to pay attention to your work and you need to

16:26

generate output, right? So in her case,

16:28

I'm just wondering like why the company would

16:30

recommend for her to take annual leave. It could be

16:32

because they might find that they are not sure whether

16:34

she needs to devote more of her time and attention

16:36

to caregiving, whether she's available

16:38

to respond when they need her to. Correct. So

16:40

on one hand, we can see as like they are rejecting her

16:43

flexible work arrangement request, but on the other

16:45

hand, they could really be just helping her

16:47

to prioritize what's really important. Like if

16:49

care for your husband is really important and you

16:51

have to be there, then maybe it's better to

16:53

devote yourself to the care, rather than try to

16:55

split yourself both ways, right? But of course, for

16:57

a lot of us, we feel like we can manage everything and that's

16:59

why we want to do the flexible work arrangement. So

17:02

if that's the case, then maybe Sandra really ought to

17:04

just have a discussion with the HR to explain

17:06

why. Like why taking leave would be

17:08

overkill, or be too much for her situation

17:10

and how she's still able to cope

17:12

with her work capacity, to

17:14

respond on time, and that her attention

17:16

would not be diluted too much if

17:19

she's providing care for her husband. Flexible

17:21

work arrangements should not be a decision,

17:23

like a judgment, right, like or it's stamped already.

17:25

I think it should be more of a discussion. Both

17:28

sides need to engage in conversation

17:30

to explain why certain things are done. Yeah. So

17:32

on that point, if the flexible working arrangement is

17:34

rejected. Sandra could perhaps

17:36

also still have a conversation with her superior

17:39

and to say, you know what, could I maybe take

17:41

some time off? Let's say my husband has to go

17:43

for a follow-up checkup, can I take some

17:45

time off on this day, and then I'll come

17:47

back and work another 2 hours or 3 hours. I

17:50

think like you say, it's always a discussion. They

17:52

might actually be open to letting you

17:54

be away from the keyboard for 2 or 3

17:56

hours and then come back and pick up the work.

17:59

So the guy is there, but we need to

18:01

learn how to communicate and converse properly.

18:03

Employees really need to know. How to put

18:05

up a case and to explain

18:07

how they can still be productive. At the same

18:09

time, employers also need to start to realize that, hey,

18:12

I'm not able to get 100%

18:14

of an employee, right? Even if today

18:16

someone goes to work sitting in front of the computer at the

18:18

workplace, you may not be getting 100% attention from that

18:20

person, right? So the employees also on

18:22

the other hand, need to know that, OK, what's the bare

18:24

minimum that I'm willing to accept, to say

18:26

that if today you're out of sight, I still know

18:28

that you're performing even though your attention might be divided.

18:31

Yeah. So this is something that I think we will

18:33

need to. Figure it out along the way, right? There's

18:36

going to be a lot of teething issues, I think

18:38

surrounding the new flexible

18:40

working arrangement guidelines and we're working on

18:42

a longer podcast so that we can

18:44

discuss this in greater detail. But

18:46

Sandra, thank you for your question and we really

18:48

hope that your husband will recover

18:51

very, very soon. If like Sandra, you

18:53

have a work-related question, do write to

18:55

us. We are at CNA podcasts

18:57

at Mediacorp.com.sg. You

18:59

can also find us on Spotify, Apple Podcasts

19:02

and YouTube. The team behind

19:04

the Work It podcast is Christina Robert,

19:06

Joan Chan, Juani Johari and

19:08

Saye Win. Sound mixing is by

19:10

Carrie Lim, video by Reza Rahman

19:12

and Hanida Amin. I'm Gerald and

19:14

I'm Tiffany. Here's wishing you a

19:16

good work week ahead.