00:00:00
One of the concerns in the labor
market right now is a perceived
00:00:04
mismatch in skills, what
companies are looking for and
00:00:07
what might be actually available
in the labor market. Part of the
00:00:10
reason why is a very quick
reorganization going on at the
00:00:14
moment, and to no surprise, AI
is playing a role in all of
00:00:18
this. But how do we best judge
the skills gap? Recently,
00:00:22
Wharton and Accenture came
together to try and tackle that
00:00:25
issue. Pleasure to be joined by
Eric Bradlow, who's Vice Dean of
00:00:28
AI and Analytics at Wharton, as
well as Professor and Chair of
00:00:32
the Wharton School of Management—
excuse me, Marketing Department.
00:00:36
Eric, great to see you again.
How are you?
00:00:38
It's great to be on
the show, Dan.
00:00:40
Thank you. All right.
I've heard the
00:00:43
conversation about skills gap
for at least a couple of years
00:00:47
now. How significant of an issue
is it?
00:00:51
Well, the best way I can
describe that it's really
00:00:54
significant is, as you can
imagine, given my role at AI and
00:00:58
Analytics at the school, I talk
to alumni. Go travel the world,
00:01:02
talk to our current students.
The minute I'm done talking, the
00:01:06
number one question asked of me
is, what— let's say it's for
00:01:11
undergraduates. What should my
high schooler be studying? What
00:01:15
jobs are going to be left over?
It's the same question asked for—
00:01:19
by undergrads, MBAs, at all
levels. I think it's the most
00:01:23
significant question from our
side, from the academic side,
00:01:27
because what should we be
training students to do? And the
00:01:32
only way to understand that— and
remember, we're Wharton, which
00:01:35
means we're the empirical
school where and we're going to
00:01:37
talk about this, we're not just
going to opine on the skills
00:01:40
gap, we're going to actually
measure the skills gap.
00:01:44
And so that's part of the reason
why this partnership between
00:01:46
Wharton and Accenture came about,
to kind of better understand
00:01:50
what's happening and how to
address it.
00:01:52
I think, let me— it's a good
question. The Wharton partnership
00:01:56
with Accenture— and I should say,
it is the flagship partnership.
00:01:59
We call it the affiliates
program. Accenture is the number
00:02:02
one affiliate of Wharton AI and
analytics. We're doing a little
00:02:06
branding here today for that.
This is one piece of a much
00:02:11
larger partnership, which also
involves developing content
00:02:16
around AI and analytics, events
around AI and analytics, thought
00:02:20
pieces around AI and analytics.
But absolutely, you know,
00:02:24
Accenture is, of course, their—
one of their big parts of their
00:02:28
business model is providing
support to companies in areas
00:02:32
where there are those gaps. And
so it's not just to enlighten
00:02:35
us, but it's for their structure
and their reorganization that
00:02:39
they want to understand what the
skills gap is as well.
00:02:42
And the term skills-based
economy seems to be growing in
00:02:46
importance right now from the
business side.
00:02:49
I think you've said the correct
word, in my view, at least four
00:02:52
times since this interview
started. Which is, it's not a
00:02:56
jobs gap. And that's the crucial
part. People ask, you know,
00:03:00
go back 30 years. What job
should I go into? That's not the
00:03:05
right question anymore. It's
what skills do I need to have?
00:03:09
And as a matter of fact, what
you're seeing on more and more
00:03:11
resumes— because I get— I see
resumes all the time, not just
00:03:14
from students, but as you know,
I have a company. All the time, I
00:03:18
see people applying. I do the
following, Dan. Obviously,
00:03:22
people tend to list their
experience. That's great. They
00:03:25
list their education, that's
great. I go directly to the
00:03:29
bottom of people's resumes,
which is typically where they
00:03:33
list their skills. That's what
everyone's hiring against today.
00:03:38
Everyone's hiring for people
with skills. And while this is
00:03:42
not part of the report, as you
know, for 20 years, one of the
00:03:45
things we've been doing at
Wharton AI and analytics is,
00:03:48
while people are students here
and to our alumni worldwide,
00:03:52
providing people skills-based
opportunity to take what they've
00:03:56
learned in the classroom and
apply those skills to actual,
00:04:00
real world projects. Because
that's what companies are
00:04:02
looking for today.
00:04:03
I saw something in the
report that also talked about
00:04:06
the fact that skills, in many
instances, are becoming more
00:04:10
important than the actual job
titles that companies are
00:04:13
looking to fill.
00:04:15
Oh, I could not agree with you
more. In other words, what
00:04:19
you're— let's see. The other thing
that's going to happen besides
00:04:22
my agreement with that
statement, what you're going to
00:04:25
see is assessments that firms
do. Look, I'll give you an
00:04:30
example with my middle son. He
was taking a skills-based job
00:04:33
recently at a private equity
firm. He literally went through,
00:04:36
and this is no exaggeration,
Dan, 13 rounds of skills-based
00:04:42
interviews to come up with, does
he know this? Does he know that?
00:04:46
Does he know this? And so I
think what you're also going to
00:04:50
see, it's going to back up just
from, do I hire Dan Loney? Do I
00:04:53
hire Eric Bradlow? You're going
to see the evaluation of what
00:04:58
skills you have increase.
And actually, the nice—
00:05:01
another thing about
artificial intelligence, by the
00:05:03
way, is you can ask it to
construct those skills-based
00:05:08
assessments. And not only will
it say, "Eric Bradlow has this
00:05:11
skill," or not, but if I don't
have that skill, it will provide
00:05:15
me the education I need to
develop that skill.
00:05:18
What does that mean, then, in
terms of the needs, the
00:05:23
necessities that the individual
will need, or the student coming
00:05:28
out of college, or the student
coming out of high school going
00:05:30
into the workforce, that they
will need to have, or at least
00:05:34
start to develop while they are
still going through their
00:05:37
education process?
00:05:39
Yeah, so you talk about two
things. Three— actually,
00:05:43
there's three answers to your
question. The first is, is that
00:05:46
Eric Bradlow and Dan Loney could
be sitting in the same classes,
00:05:49
taking the same curriculum, but
because of artificial
00:05:53
intelligence and the degree and
the ability of personalization,
00:05:56
Dan will get— even if we have
the same grades, Dan will get,
00:06:01
"Your skills are deficient in
this area." "Eric, your skills are
00:06:04
deficient in this area." The
second thing is, Dan will get,
00:06:08
"Here's your individualized
lesson plan to build your skills
00:06:12
in this area." "Eric Bradlow, here
are yours. Here's your lesson
00:06:15
plan to get skills in this area."
The third thing is— and this, Dan,
00:06:19
is going to take a little bit
longer, because we in education
00:06:23
tend to not be the most
innovative and fast moving. How
00:06:27
does the curriculum change from
course labels to skill-based
00:06:33
labels? And that's what— by the
way, that's what students know.
00:06:36
Because students, by the way—
students typically lead us
00:06:39
faculty. Students know the
skills they need, and they're
00:06:43
going to be looking for courses,
whether it's whether we try to
00:06:45
tell it in the title of the
course or in the description,
00:06:49
trust me, students— here's what
I know students are doing, Dan.
00:06:52
You'll like this. They're
ingesting syllabi from faculty
00:06:56
about courses and seeing whether
it matches the skills they want
00:07:00
to acquire.
00:07:02
Are, then, the changes
that we will see occur in the
00:07:05
education side, are they more
subtle in terms of the
00:07:11
development off of what is
already being taught, or are we
00:07:15
talking about a complete 180?
00:07:18
Yeah, it's a great question. So
00:07:20
there'll be differing of
opinions on this, but since I'm
00:07:23
here, I get to give my opinion
on this. - Right. Right.
00:07:25
I think everything has to change.
I'll just give you an example of
00:07:29
myself, and I know I speak not
just for myself. One of my
00:07:32
colleagues who works with us at
Wharton AI and Analytics started
00:07:36
what's called the Wharton
Generative AI Lab, which is
00:07:38
Ethan Mollick, who's extremely
well known now, is well-read
00:07:42
and well-posted on these
topics. He and I both believe
00:07:45
that pretty much everything has
to change. Let me give you an
00:07:48
example. So I teach— I'm about
to start, literally two weeks
00:07:51
from today or two weeks from
yesterday. I teach our MBA Core
00:07:54
here at the Wharton School. And
it's, of course, not
00:07:56
surprisingly, Dan, in data-
driven marketing, which brings
00:07:59
together my love of data science
and marketing. Up until two
00:08:03
years ago, this was a course
where I would have students
00:08:06
ingest data, analyze data using
R or Python, come out with
00:08:11
analyzes, interpret those
analyzes and say what it means
00:08:14
to the firm. Well, now, of
course, the generation of data
00:08:18
can be done using generative AI,
whether it's from real data or
00:08:22
simulated data. The analysis is
now about prompt engineering.
00:08:26
It's not about teaching them R or
Python, because you can code
00:08:29
that directly in large language
models. So literally, every
00:08:34
aspect of my course, from
evaluation— I have Dan Loney
00:08:38
read a case. Let's be honest.
Dan Loney is taking that case,
00:08:41
ingesting into a large language
model, and is going to ask it for
00:08:44
the answer. A matter of fact, if
Dan Loney is a smart man, Dan
00:08:47
Loney is doing that. And by the way,
that's the skills that jobs want
00:08:52
you to have. They're not going
to drop 50 PDFs on your thing
00:08:55
and say, "Print them out and read
them." They know you're not going
00:08:58
to do that. So I'm a believer
that we need to totally rethink
00:09:03
the way we educate people around
skills. And I know we're going
00:09:06
to talk in a second about which
skills I think tend to be over
00:09:10
versus undervalued.
00:09:11
Well, that was going to be my
next question, because from a
00:09:14
business perspective, they have
to look at this from the bottom
00:09:17
line, which means the financial
side of this. So there is a
00:09:21
value when you're talking about
what skills are most valuable
00:09:24
and what skills are not. And can
you can you already see certain
00:09:29
skills that are kind of
universally more valuable, kind
00:09:32
of across the business, no
matter who it is?
00:09:35
So there's two answers to
your question. First, we have to
00:09:39
talk about the skills that are
valued, and then we have to talk
00:09:42
about supply and demand. There
could be a skill that's
00:09:45
extremely valued, but there's
too much of it in the market. So
00:09:48
let me give you an example. This
is why we work with Accenture
00:09:52
on this skills gap. So let me
just be clear on what we did, to
00:09:55
every— to all of our listeners
here. There's a company called
00:09:58
Light Start. What we did
is that we scraped data from
00:10:05
both job postings and from
resumes. That makes a lot
00:10:09
of sense. Job postings list—
forget just the titles— the
00:10:13
skills they— and experience they
want people to have. Resumes are
00:10:16
people highlighting the skills
and experience they have. Let's
00:10:19
see if there's a gap between the
two. That was the intent of the
00:10:23
project, to have an empirically-
driven measure that we could do
00:10:27
every year. Wharton, partnered
with Accenture, on the skills
00:10:31
gap. Now the next natural
question— remember also, jobs
00:10:34
come with salaries, which means—
or ranges, which means you can
00:10:39
now say which skills tend to be
paid more, right? You can do
00:10:43
that. You can also say
which resumes and which skills
00:10:47
in those resumes tend to get
accepted for higher paying jobs.
00:10:51
You can start doing that. So all
the things that you might want
00:10:55
to answer, you can do through
this data. So let me tell you
00:10:58
the two— the two big highlights
that came out for me. So one of
00:11:03
the things I hear all the time,
Dan, I'm sure you've heard this,
00:11:05
is that five years ago, every
child, man, woman, etc, was told,
00:11:11
"You got to study computer
science, because Big Data
00:11:14
science is the future." Now, all
I hear is, "Should my kids take
00:11:18
any coding classes whatsoever?
Because isn't computer science
00:11:22
going to be entirely
democratized?" So one of the
00:11:25
skills we found is that there's
a oversupply of people with
00:11:30
claimed computer science skills.
Now I'm not saying people don't
00:11:33
need those skills, but they— it's—
there's an oversupply of it. The
00:11:37
part on the other side, though,
the two things that never go
00:11:41
away— and I'm not sure AI is
going to replace those skills.
00:11:44
Maybe they will— are what I would
call communication and selling.
00:11:48
Because what happens is, if AI
can do the technical parts of
00:11:52
jobs, or parts of the technical
parts, you still need to be able
00:11:55
to sell the ideas. You still
need to be able to sell your
00:11:58
company services. So that part,
there's an under— let me say, be
00:12:02
clear— there's a claimed
undersupply, which means maybe
00:12:06
people aren't highlighting the
right skills in their resumes.
00:12:10
The second thing that
came about was— you know, I hate
00:12:13
to say this, but too many people
claim, "I've got leadership
00:12:17
skills." Well, the problem— I'm
not saying leadership skills
00:12:21
aren't important. All I'm saying
is, when we look at the supply
00:12:25
and the demand of leadership
skills, there are too many
00:12:28
people that say they have
leadership skills compared to
00:12:31
jobs that are hiring people that
say, "We need someone with
00:12:35
leadership skills."
00:12:36
What do you think, then, this
means for the labor market as we
00:12:39
go into the future?
00:12:40
Well, let me— that's good.
Great question. These are great
00:12:43
questions. That's why I love
talking to you, Dan.
00:12:45
Thank you, Eric.
00:12:46
No, no, no. This is why
we've been together for— we've
00:12:48
been partnering for 15-plus
years on this. - Oh my god, yeah.
00:12:52
It's been— and then sometimes it's
around sports, sometimes it's
00:12:54
around <i>This Week in Business</i>,
it's all kinds of things. - Yep.
00:12:57
I would say two things. Here's
what's going to happen in the
00:12:59
short term. I think, obviously,
in my view, the Wharton School,
00:13:03
Accenture, we're going to push
this report out. Big launch
00:13:06
tomorrow. Big, big PR launch
about this. First, if I were a
00:13:10
rational person looking for a
job in the short term, maybe I
00:13:14
can't do skill acquisition, but
I'll put on my marketing hat. I
00:13:18
can reposition the skills I have
around the skills that companies
00:13:23
are demanding. So rather than
emphasizing my leadership
00:13:26
skills, which I'm not saying you
don't need to have, but maybe
00:13:30
you should emphasize the things
that we're finding.
00:13:33
Communication, your technical
skills, because those are the
00:13:36
skills that there seem to be an
undersupply for. So I think this
00:13:39
is literally marketing. What
you're going to see is a short
00:13:43
term repositioning of people in
the way they describe their
00:13:47
skills. Now in the medium term,
this is where, whether it's
00:13:51
academic institutions with full
time jobs or, what it is, it's
00:13:56
other institutions that are
actually coming up with short
00:14:00
term courses. That's what we're
going to see. We're actually
00:14:04
going to see people saying, "I
see there's a shortage of
00:14:08
digital marketing skills.
There's a shortage of web
00:14:10
development skills." They're going
to come in and create that content.
00:14:15
Eric, always great to get your
thoughts and your insight.
00:14:18
Thanks again. Look forward to
talking again down the road.
00:14:21
It's great to talk to you, Dan,
and I really believe this
00:14:24
Wharton Accenture Skills Gap
Report will be something we hope
00:14:28
that is tracked as much as
anything over time.
00:14:31
Eric Bradlow, Vice Dean of AI
and Analytics
00:14:34
here at the
Wharton School.
