00:00:00
Well, great to have back with us Wharton
00:00:02
marketing professor Stephano Pantoni as
00:00:04
we talk about AI. Stephano is also
00:00:07
co-director of Wharton's human AI
00:00:09
research. Uh our conversation this month
00:00:12
is about a new report that they have put
00:00:14
out looking at how companies are
00:00:16
implementing generative AI and how they
00:00:18
are looking to gain an advantage on
00:00:21
their competition. Stephano, great to
00:00:23
talk to you again. How are you sir?
00:00:25
>> Thanks Dan. Great to great to be here.
00:00:27
>> And so this is a continuing study. This
00:00:29
has been going on for a couple years as
00:00:31
generative AI has kind of continued to
00:00:33
develop.
00:00:34
>> Yeah, we uh we're in the third year now.
00:00:37
We basically discussed with our partner
00:00:40
GBK Collective, a market research
00:00:42
company. Basically just a matter of
00:00:43
weeks after uh CHP was released in
00:00:46
November 2022 and by summer 2023 we were
00:00:50
out in the field collecting data. We did
00:00:52
that in 2024 and we did it again this
00:00:55
year. So now we've got three years of
00:00:56
data which is nice. Where are we in
00:00:59
terms of the development of generative
00:01:01
AI and kind of the understanding that
00:01:03
that companies and maybe more
00:01:05
importantly seuite executives have on
00:01:08
how much it is going to be used.
00:01:10
>> Yeah. So the survey it's a large sample
00:01:13
of senior executives and management
00:01:16
directors and about half of the sample
00:01:18
are vice president or seuite level. So
00:01:20
it's a it's a a way of getting the pulse
00:01:23
of senior leadership in large US
00:01:26
companies. That's basically our sampling
00:01:28
frame and we measure lots of questions.
00:01:30
We ask them a lot of question about
00:01:31
their own use of generative TVI and
00:01:34
their organization and their functions.
00:01:35
We have people across basically many
00:01:37
industries and all functions and so we
00:01:39
are getting bit of a picture or where
00:01:42
companies stand in their geni adoption
00:01:44
process.
00:01:45
>> Obviously a lot of questions around gen
00:01:48
AI and one which seemingly draws a lot
00:01:50
of attention is how it will have an
00:01:52
impact on labor. uh what do the
00:01:55
executives feel like is the the path
00:01:59
that we're going to see Gen AI and the
00:02:02
impact it'll have because it's
00:02:03
interesting you note in the report about
00:02:05
entry level jobs and internships and how
00:02:08
those may not be as impacted as much.
00:02:11
>> Yeah. So the conversation on jobs about
00:02:13
geni has been uh quite uh um you know
00:02:17
scary or gloomy for already a while and
00:02:19
I think when you get in conversation
00:02:21
with people about AI and jobs you end up
00:02:23
talking about Terminator and the Matrix
00:02:24
in five minutes. So you know there's
00:02:27
that kind of vibe and especially it's
00:02:29
been gloomy around entry- level
00:02:31
positions. people have talked about uh
00:02:33
the end of entry- levelvel positions
00:02:35
anywhere that basically now you can have
00:02:36
those AI interns who do a lot of the
00:02:39
work that maybe you get a summer intern
00:02:40
on MBA graduate or something to do uh
00:02:43
now using tools like deep research or
00:02:46
other geni tools and so uh obviously um
00:02:50
geni is going to have a big impact on
00:02:52
entry-level positions because the people
00:02:54
in the entry- level positions tend to be
00:02:56
the people who do the work right the
00:02:58
people above them tell people what to do
00:03:00
but they at the entry will actually do
00:03:02
it. So obviously they're going to be the
00:03:03
ones who have to navigate and negotiate
00:03:05
the impact of Genai tools on uh
00:03:09
workflows and functions and so on. Now
00:03:11
the interesting thing in our data is
00:03:13
that when you hear um the answers of the
00:03:17
senior leaders, you get a picture of a
00:03:19
technology that is powerful, everybody
00:03:21
agrees is going to have a major impact
00:03:23
on every function. So it's not hyped,
00:03:25
it's it's real. And at the same time
00:03:28
though they tend to tell us that while
00:03:30
human replacement concerns are high in
00:03:33
fact people as a net effect think that
00:03:36
the technology is going to have mostly
00:03:38
the effect of augmenting or enhancing
00:03:41
employee skills. So in a way they think
00:03:43
that the technology is powerful but it's
00:03:45
not quite doing what a human expert is
00:03:48
doing and so we are hoping to see
00:03:49
complimentarities between human
00:03:51
expertise and AI. Now at the entry level
00:03:55
that's where the conversation has been
00:03:56
very negative. In our data we do see
00:03:58
that people tell us that uh the negative
00:04:02
impact on jobs is going to fall
00:04:04
disproportionately
00:04:06
on the junior positions relative to the
00:04:08
senior position. They do say that. But
00:04:11
what is missing from the conversation
00:04:13
generally is another finding that we
00:04:16
also get in the data which is while they
00:04:18
say that the negative impact will fall
00:04:21
disproportionately on junior people,
00:04:22
they also say that the positive impact
00:04:26
will also fall disproportionately on the
00:04:29
young people. And the net effect is not
00:04:31
so obvious. Maybe in fact if you look at
00:04:32
the the numbers in our data indirectly
00:04:34
it might be actually a good thing for
00:04:35
ent certainly those jobs are changing
00:04:37
the most. Why is it that that bo both
00:04:41
you will see the negative and the
00:04:42
positive uh in the entry level category?
00:04:45
>> Because it's a disruptive technology.
00:04:46
You know, these jobs are going to change
00:04:48
a lot and so and new jobs are going to
00:04:50
come and new ways of doing things are
00:04:51
going to come up and uh people are going
00:04:52
to be able to perform things and now and
00:04:55
require much more labor and maybe you
00:04:58
know specialized expertise. So lots of
00:05:00
jobs are really going to be different.
00:05:02
Uh but that doesn't necessarily mean you
00:05:05
know need to mean that the jobs are
00:05:07
going to disappear.
00:05:09
So from the business perspective, we are
00:05:11
seeing obviously bigger and bigger
00:05:12
investments made by companies in and
00:05:14
around uh AI. How much of this is
00:05:18
basically a race against their
00:05:20
competition right now? Yeah, there's a
00:05:23
lot of FOMO I think where a lot of firms
00:05:25
feel they have to spend money on Genai
00:05:28
uh because everybody's doing it or
00:05:29
because if they don't do they look like
00:05:31
they're antiquated or whatever and uh
00:05:33
there is a danger in that in so far as
00:05:36
you know maybe some peer kind of uh uh
00:05:40
envy of that kind
00:05:44
for people to change. At the same time,
00:05:46
what you don't want to do is to do
00:05:47
something only because everybody else is
00:05:48
doing and then you end up with this kind
00:05:50
of a herd mentality and that's not
00:05:51
helpful. I think the important thing is
00:05:53
for companies to realize this technology
00:05:55
has very impressive capabilities and
00:05:57
then think very very hard about how
00:06:00
those capabilities can be used to do a
00:06:04
better business and that I think is what
00:06:06
we don't know we are learning.
00:06:08
>> What's the expectation then that of the
00:06:10
type of impact that geni will have? You
00:06:13
talk about the fact in the report this
00:06:15
latest one that it won't necessarily be
00:06:18
a narrow impact but it could very well
00:06:20
be a pretty broad-based one.
00:06:22
>> Yeah, we do we have lots of interesting
00:06:24
question to speak to that and one of the
00:06:26
section of the surveys about use cases
00:06:28
and we see companies endorsing very
00:06:29
large number of use cases already making
00:06:31
progress on most of them. um general
00:06:33
office work um is going to be impacted a
00:06:36
lot but also a lot of uh functional
00:06:38
applications like say marketing could be
00:06:39
customer service versus marketing
00:06:41
content creation things like that now
00:06:43
what's interesting to me that this year
00:06:45
we added for the first time some
00:06:47
measures on return on investments which
00:06:49
is a big question for companies are we
00:06:51
getting the returns we were hoping to
00:06:53
get out of these genai investments and
00:06:55
the mood there has also been quite
00:06:57
negative over the last months or two lot
00:06:59
conversations seem to suggest that
00:07:01
companies are struggling to see positive
00:07:04
returns some anecdotal evidence of that
00:07:06
and in our report actually we don't see
00:07:08
that what we find is that 3/4 of the
00:07:11
people who say they're measuring ROI
00:07:13
which is the majority of the firms are
00:07:15
reporting positive ROI modily or
00:07:17
significantly positive
00:07:19
>> so is that because of how the companies
00:07:22
are using AI how it's being implemented
00:07:25
and and in different areas of their
00:07:27
business they they've kind of taken the
00:07:29
the belief that we can use this in so
00:07:32
many different areas that we can get a
00:07:34
better ROA right now.
00:07:36
>> Yeah. And you know, these are
00:07:38
self-report survey data. So, you should
00:07:40
always take uh the numbers, exact
00:07:43
numbers with a pinch of salt because
00:07:44
this is what they tell us. But I don't
00:07:46
have a view of the hard data inside the
00:07:49
companies. You know, they might be over
00:07:51
reporting or maybe you're too optimistic
00:07:54
about the results or they might not even
00:07:56
know and taking a guess. So, there might
00:07:57
be a lot of bias in these numbers. It is
00:07:59
possible. These are self-reports. is a
00:08:01
caveat that we always need to
00:08:03
acknowledge with this kind of data. But
00:08:05
I do think that it gives us a very good
00:08:07
sense for the vibe, you know, how people
00:08:09
feel about the technology and the vibe
00:08:11
is good. The vibe is that this
00:08:13
technology is delivering positive
00:08:15
results.
00:08:15
>> Finally, what are then the future
00:08:18
expectations, you know, 10, 15, 20 years
00:08:21
down the road in terms of the impacts of
00:08:23
AI?
00:08:24
>> Oh, we can't we can't speak to 10, 20
00:08:26
years, 30 years is impossible. I I I
00:08:28
always say I never make predictions
00:08:29
beyond 24 months but uh uh I think what
00:08:32
we're seeing is companies are continuing
00:08:34
to invest. Um but there are some
00:08:37
tensions and the tensions tend to be not
00:08:39
so much on the capital spending or the
00:08:41
um or the technology itself seem to be
00:08:44
more on the human capital side. So
00:08:46
companies report on difficulties in
00:08:48
finding the right talent in upskilling
00:08:50
and hiring and also some beliefs that
00:08:52
the training is not really working as
00:08:54
well as it should. There's a lot of
00:08:56
interesting insights about that in the
00:08:57
report too. But basically, I think that
00:08:59
uh geni adoption is a big challenge for
00:09:02
organizations. We don't have a template.
00:09:04
We have to figure it out. And that
00:09:05
challenge is actually more of a human
00:09:07
capital challenge than a tech challenge.
00:09:09
>> Stephano, great to talk to you once
00:09:11
again. We will catch up with you again
00:09:13
in a month and talk more about the world
00:09:15
of AI.
00:09:16
>> Great. Thanks, D.
00:09:17
>> Thank you. Stephano Pantoni who is a
00:09:19
marketing professor here at the Wharton
00:09:21
School and co-director of Wharton Human
00:09:23
AI research.
