AI Won't Replace Humans—But Humans With AI Will Replace Humans Without AI
Harvard Business Review2023-08-03
23K views|11 months ago
💫 Short Summary
Karim Lakhani discusses the impact of AI and machine learning on business, emphasizing the shift to an AI-first world and the need for companies to adapt to new technologies for success. He highlights the importance of continuous learning, addressing biases in AI systems, and embracing generative AI. The discussion also touches on the evolving relationship between humans and AI, potential advancements in natural intelligence, and the possibility of encountering alien life forms. Overall, the message encourages embracing technological advancements and exploring new ideas in the changing business landscape.
✨ Highlights
📊 Transcript
✦
Impact of machine learning on the rules of business.
02:25Technologies like AI are fundamentally changing the nature of corporations.
Collaboration with Marco Iansiti and Amy Bernstein on the book 'Competing In the Age of AI' based on a decade of research.
Insights gained from working with various companies on the transformation of the modern American corporation.
Advancements in AI and machine learning are reshaping the corporate landscape established in the 1920s and '30s.
✦
The impact of AI-first companies on business architectures and customer experiences.
04:44Machine learning and AI are fundamentally changing business operations and value delivery.
Companies like Google and Netflix are leading the way in automation and AI integration.
Traditional companies, such as General Electric, operate differently from those centered around machines and algorithms.
To meet consumer expectations, most companies will need to adopt AI and digital technologies for seamless experiences.
✦
Transition to an AI-first world is inevitable, with decreasing costs and a playbook for implementation.
07:43Main challenge is organizational, requiring a digital mindset among executives and workers.
Companies need to provide top-notch user experiences akin to Google and Amazon to meet customer expectations.
Future technology waves like generative AI and quantum computing will require quick adaptation and a culture prepared for advancements.
Executives must prioritize continuous learning to stay competitive in the evolving digital landscape.
✦
Importance of Learning Accounting, Digital Technologies, and Machine Learning in Business.
11:23Continuous learning is crucial for personal and professional development.
Companies should invest in employee learning and adapt to change management.
Change management is a vital skill for leaders and workers to navigate the evolving business landscape.
Embracing new technologies, adapting to change, and staying current are essential for success in today's world.
✦
Reflection on challenges faced with technology during COVID.
15:05Difficulty of crossing borders with QR codes and apps.
Comparing executives and companies struggling with technology adoption to elderly individuals needing help with new tools.
Discussion on generative AI and its various applications.
Emphasis on embracing technological advancements for progress.
✦
Impact of Internet and Generative AI on Industries
19:52The internet lowered information transmission costs, enabling companies like Google, Amazon, and Facebook to rise.
Generative AI is reducing the cost of cognition, changing how we think and work.
Researchers are exploring the potential of generative AI and its impact on different industries.
The shift from the internet to AI deployment signifies a transformative era reshaping industries and daily life.
✦
Importance of Embracing Generative AI in Businesses.
22:33Companies are encouraged to start using generative AI and not resist its implementation.
Generative AI can be utilized for content generation and personal assistance.
The AI field is experiencing rapid innovation and application development.
Leaders and managers are urged to adopt and explore the potential of generative AI instead of fearing or restricting its use.
✦
Importance of running boot camps for everyone and providing access to tools and ranking based on use cases.
25:09Publishers' reluctance to accept generative AI-produced content, highlighting the need for authors to take responsibility.
Discussion on the attribution dilemma in academic writing and the challenges of biases in AI and machine learning.
Highlighting Mozilla's initiative to create open-source language models to detect and address biases as a potential solution.
✦
Importance of Addressing Bias in AI Systems.
28:27Understanding existing biases in society is crucial for mitigating bias in machine learning models.
Representative data, training, and labeling are key components in addressing bias in AI systems.
Leaders have an ethical responsibility to identify and correct biases to avoid legal consequences.
AI has the potential to either amplify or correct existing biases, depending on one's perspective as a techno optimist or pessimist.
✦
The importance of human interaction with AI and the evolving relationship between the two.
32:44Current human-like responses from AI are a statistical illusion, trained by humans to mimic human behavior.
The concept of strong AI, which is predicted to arrive within 20 years.
The merging of biology and AI is discussed, highlighting the potential for advancements in natural intelligence.
The dialogue emphasizes the need to be kind to robots and treat them with respect.
✦
Exploring the possibility of alien life forms with unique characteristics.
35:21Reference to a paper titled 'First Contact' by a machine learning expert at Microsoft.
Anticipation of Congressional hearings on aliens and the need for preparation.
Message of gratitude to the audience and sponsors for exploring new ideas in the changing work landscape.
00:00[MUSIC PLAYING]
00:13ADI IGNATIUS: Hi, and welcome
to Harvard Business Review's
00:15The New World of Work.
00:16I'm Adi Ignatius,
Editor In Chief of HBR.
00:19And each week on this show,
I interview a CEO, a thought
00:22leader, or somebody
else interesting
00:24who can inspire us and
educate us on the changing
00:26dynamics of the workplace.
00:28Whether you're navigating
the complexities
00:30of a large corporation or the
challenges of a-- excuse me--
00:34of a small startup,
whether you're
00:35based in the US or
anywhere else in the world,
00:38you face your own
unique set of issues.
00:41So the aim of this podcast is
to inspire thought and provide
00:44insights as you seek to
bolster your business
00:46and pave your own path
toward career success.
00:49So on today's episode, we have
a great guest, Karim Lakhani
00:52of Harvard Business School.
00:53I'm going to come back with
a proper introduction in just
00:55a moment.
00:56But first, let's hear
from our good friends
00:58at KPMG, who are sponsoring
this season of The New
01:01World of Work.
01:08ANNOUNCER: At KPMG,
it's our people
01:10who make the difference for
our clients, talented teams
01:13leveraging the right
technology to uncover insights
01:16that illuminate opportunity.
01:18Ready to make the
difference together?
01:25ADI IGNATIUS: All right,
just a couple more things
01:27before we start.
01:27If you're an HBR subscriber
and you're watching this,
01:30you can hbr.org/newsletters
to sign up for The New World
01:34of Work newsletter, where I try
to offer an inside look at each
01:37of these interviews and talk
about some of the ideas that
01:40come out of them.
01:41And if you like content
like this, please
01:43subscribe to our
magazine and website.
01:45The address is
hbr.org/subscriptions.
01:48And if you like hearing
smart people talk about some
01:51of these same issues, be sure to
check out our flagship podcast,
01:55IdeaCast, available wherever
you get your podcasts.
01:59And lastly, remember, you
can watch previous episodes
02:01of this show on YouTube or right
here on LinkedIn and Facebook.
02:05So let's get on with it.
02:07My guest this week, as I
said, is Karim Lakhani,
02:10a professor at Harvard
Business School who
02:12specializes in workplace
technology and particularly AI.
02:16He's done pioneering
work in identifying
02:18how digital transformation has
remade the world of business.
02:22And he's the co-author of the
2020 book, which HBR published,
02:25Competing In the Age of AI.
02:27So Karim, welcome to the show.
02:29KARIM R. LAKHANI: So glad to
be here with you today, Adi.
02:31Thank you for the invitation.
02:32ADI IGNATIUS: So there's
a lot to talk about,
02:34and I definitely want to talk
a lot about generative AI.
02:37But maybe we'll--
once we get to it,
02:39I don't think we'll get off.
02:41KARIM R. LAKHANI: No, exactly.
02:41ADI IGNATIUS: So
let's start elsewhere.
02:43But so you co-wrote a piece
for us a few years ago
02:47and it's reflected in
your book, where you say,
02:49machine learning has
basically changed
02:52the very rules of business.
02:53So that's a big statement.
02:54Tell us a little bit about
what you mean by that.
02:56KARIM R. LAKHANI: Yeah.
02:57So look, it's been a
real pleasure working
02:59with HBR and your
editors for the last 15
03:02years of my academic career.
03:05And the book really
was a partnership
03:08between Marco Iansiti and Amy
Bernstein, one of the editors
03:13at HBR, as we created this book.
03:15And what Marco and I noticed
in about a decade's worth
03:19of research and spending time
with companies, both writing
03:22cases, as advisors,
as consultants,
03:25and so forth, was that the
nature of the corporation,
03:29which really was established--
the modern American
03:32corporation, which
became the blueprint
03:34for the modern international
corporation established
03:36in the 1920s and '30s, was
changing foundationally
03:40because of technologies like
AI, like machine learning.
03:45And what we observed was
that the entire business
03:48architecture in many of these
AI first companies at the time,
03:53in terms of business model,
how you create value, how you
03:57capture value, and
your operating model,
03:59how you deliver value,
how you achieve scope,
04:02the number of
customers you serve,
04:04the number of products you have,
scale, the number of customers
04:08you serve, and learning,
these fundamental parts
04:10of a business
architecture were being
04:12rewired because of
machine learning and AI
04:15and digital technologies.
04:16And so if you just
sort of reflect
04:19for a bit on your
experience using Google,
04:23for example, much of your Google
experience is fully automated,
04:28from the ads you
see to the search
04:32you do to, if
you're using Gmail,
04:34how you interact with them.
04:36And so it's not people
that do those activities.
04:39It's the algorithms
that make that happen.
04:42Similarly, your
experiences with all
04:44the large e-commerce platforms,
like Amazon or Alibaba,
04:49let's say, in China,
if you imagine
04:52what happens at Netflix,
for example, again, all
04:56these examples,
we've been using.
04:58But these companies work in
a fundamentally different way
05:01than a company like
General Electric, where
05:03I grew up as a right out
of college in my first job
05:07were set up.
05:09And these companies,
the machines
05:11and the algorithms at the
center, the work is automated.
05:15The humans are actually
designing the algorithms
05:17and testing them
and checking them,
05:19making sure they're
working within bounds,
05:21but the actual
transactions and activities
05:23are being mediated
through the machines.
05:25ADI IGNATIUS: So
I'm guessing that--
05:27first of all, if
you're watching this,
05:28if you have your own
questions for Karim
05:32about digital transformation,
about generative AI,
05:34which we'll be talking
about in a moment,
05:36please put them into
the chat and I'll
05:38try to get to audience
questions later.
05:40So that was helpful,
Karim, but I
05:43would guess some of our
viewers are listening to you
05:46and thinking, yeah,
that's what we've done.
05:48And others are thinking, I'm
not sure we're on that journey
05:51or we're far enough
along on that journey.
05:53So my question is, what's
your advice to people
05:56who are listening to this
who are like, yeah, I get it.
05:59I get that there's value here.
06:00I'm not quite sure if my
company is right for this
06:03or I'm not quite sure
what the next steps are.
06:05KARIM R. LAKHANI:
Yeah, absolutely.
06:07Look, first I would say is that
I think most companies will not
06:10have a choice but to
adopt AI and to adopt
06:12digital at the core functions.
06:14In many ways, your
personal lives as
06:18mediated through
your transactions
06:19through your smartphone,
through these devices,
06:24and how you interact with
consumer technology products
06:29and so on and so forth, you're
already living in an AI age.
06:32And the thing I would say, Adi,
which is so interesting to me,
06:35is-- and I learned this
from some conversations
06:37I had with folks in India.
06:39They said, we have
some folks who get mad
06:43when your Uber car
or your Ola car
06:46or your DiDi car or your GrabCar
doesn't show up in 3 minutes.
06:51You want this magical
taxi experience.
06:53You go on your app, you press
the thing, boom, it shows up.
06:56And if it's going to be 5 or 7
minutes, you kind of get mad.
07:00And I'm reflecting on when I
first moved to Boston in 1997,
07:04and it would take me a week
to book a taxi in Boston.
07:08And so now we get mad.
07:10And similarly, if
there's a transaction
07:11dispute on Amazon or on Uber,
automatically solved, done.
07:15But the same people,
the same executives,
07:18in their own companies
are completely
07:20satisfied if a customer
service interaction
07:23can take two weeks, if
onboarding a new vendor
07:26takes six months.
07:28And so we're living in
this disconnected world
07:31where most people,
most consumers
07:34are living in this AI first
world, in their experiences,
07:37with many of these platforms.
07:38And then they
encounter our companies
07:40and our organizations, and
they're like, what is this.
07:43And so my sense is that
this is inevitable.
07:46This transition is
really inevitable.
07:49And for the folks
that are behind,
07:54the good news is that the
cost to make the transition
08:00keeps getting lower and lower.
08:03The playbook for this
is now well known.
08:07And finally, the
real challenge is not
08:09a technological challenge.
08:11I would say, that's
a 30% challenge.
08:14The real challenge is 70%, which
is an organizational challenge.
08:18My great colleague,
Tsedal Neeley,
08:20talks about the digital mindset.
08:22Every executive,
every worker needs
08:25to have a digital
mindset, which means
08:27understanding how these
technologies work,
08:28but also understanding
the deployment of them
08:30and then the change
processes you
08:32need to do in terms
of your organization
08:34to make use of them.
08:36ADI IGNATIUS: It's
really interesting
08:38because we think about this.
08:39We're a relatively
small publisher
08:42compared to some of
the giants out there.
08:45But when people come
to our site and they're
08:47searching for articles
by Karim Lakhani,
08:49they're used to a Google
sort of search experience.
08:52KARIM R. LAKHANI: Exactly.
08:53ADI IGNATIUS: When they want
to buy a product from us,
08:55they're used to an Amazon.
08:57And anything short of that,
it's like your Uber example.
09:00There's a frustration
and expectation.
09:02So we have to find ways to lift
our game without the resources,
09:05whether it's through
partnership or other things,
09:08because it's table stakes.
09:09People just expect
the best experience
09:12in every experience they have.
09:14KARIM R. LAKHANI: 100%.
09:15If I look at my
teenage daughter,
09:18she has no patience for
[INAUDIBLE] companies.
09:21And she just gets mad and
just is, like, what is this.
09:25And so absolutely.
09:27ADI IGNATIUS: So anyway, so
the next big wave-- again,
09:30we're going to talk about
it is, I keep teasing it,
09:32but it's generative AI.
09:33But that won't be the last wave.
09:35And quantum will
hit us at some point
09:37and things we don't even can't
even anticipate will hit us.
09:39So how do you prepare for that?
09:43How do you create a
kind of, I don't know,
09:45culture or mindset
or organization that
09:50knows that there will be
unexpected ways of technology.
09:53They'll come we'll have
to figure out if they're
09:56relevant to us or not.
09:57And if they are, we
need to adapt quickly.
10:00Is there a general way to
think about that [INAUDIBLE]??
10:03KARIM R. LAKHANI: Yeah, yeah.
10:04Yeah.
10:04Thank you for that question.
10:05And in fact, I've been
pondering this quite a bit.
10:08I think there are two
imperatives for most
10:11executives, for most
managers, for most leaders.
10:14One is a learning imperative.
10:17This is, again, Tsedal's work
on digital mindset, my work.
10:20There's lots of
learning you need to do
10:21and the learning has
to be continuous.
10:23And the idea is that not that we
want you to become AI engineers
10:27or data scientists or get a PhD
from Stanford or Harvard or MIT
10:32or Tsinghua University or Oxford
University, or from the IITs
10:38in machine learning.
10:39But as executives, this
is now table stakes.
10:44And the way I think about
this for at the MBA program,
10:47for us, people come to
the Harvard MBA program.
10:52And we have-- first year
is a required curriculum.
10:55There's 10 courses, and
one of them is accounting.
10:58Now, I can tell you, accounting
is a very important profession.
11:01But most MBAs that join HBS
don't want to be accountants.
11:05But they need to learn
accounting because that's
11:07the language of business.
11:09That's the ways
in which you think
11:10about how value
is kept track of,
11:13how expenses are being tracked
of, and so on and so forth,
11:15super important.
11:16And you don't take
the accounting course
11:18to become an
accountant, but you need
11:20to understand accounting so you
can be a good business person.
11:23Same thing now with digital
technologies and machine
11:25learning.
11:26You need to understand
the machine learning
11:28stuff and the AI stuff,
not because you're
11:31going to become an engineer
or an AI scientist,
11:34but because that is now
going to be a critical table
11:36stakes for you to understand
how business works.
11:38So there's a learning
imperative and I
11:41don't think we can take away
the learning imperative anymore.
11:44Now, look, I'm
self-centered about this.
11:46I'm self-interested.
11:46I'm in the learning profession.
11:48That's what I do, so
I want to caveat that.
11:50But I want to just insist that.
11:52I think the learning
journey does not stop.
11:54And you have to invest in
your own personal learning,
11:56and I think companies
need to invest
11:58in the learning for their
own employees as well.
12:00It's a two-phased conversation.
12:03Companies have to embrace
this and so do individuals.
12:07But the second
bit, I think, Adi,
12:09is equally important, which I
think is completely underrated,
12:14which is change and
change management.
12:18And change becomes a skill
for managers and for leaders
12:21and for executives.
12:22How you change, how
you continue to change,
12:25how you build the
DNA for changing
12:28becomes very important.
12:29At a time, by chance, I
was in Asia a month ago
12:33and had a chance to spend
some time with Mickey
12:35Mikitani at Rakuten.
12:36And one of the
things that's amazing
12:40is what he has thought
about as change
12:42as a core competency for all
workers and for all employees.
12:46Right now, most change programs
are viewed with skepticism,
12:49as flavor of the month, blah,
blah, blah, and people resist.
12:53People resist.
12:55I think the best companies will
be the ones that can understand
12:58how change becomes a skill.
13:00And if you think about change
as a skill, what does that mean?
13:03Skills require
acquisition of the skills.
13:05You've got to
invest in learning.
13:06What does it mean to change?
13:08It requires practice.
13:09You've got to keep
changing as well.
13:11And it requires adjustment.
13:13Once you've learned
how to change,
13:14how do you change that-- how do
you project that to everybody
13:16else?
13:17So those elements,
I think, will become
13:19a key part of the
ways in which leaders
13:25need to adapt to this world.
13:27ADI IGNATIUS: Yeah.
13:28I think that's great.
13:29It could be generational and
it should not be generational.
13:31I always think that
when you come of age,
13:35when you join the
workforce, there's
13:36a certain suite of
technology that it's
13:38just you grew up with it and
you're comfortable with it
13:40and you are part of
figuring out how to use it.
13:44And then subsequent ways, a
lot of us hit a point where,
13:47yeah, that one seems stupid.
13:49And for my father, who's still
alive at 102, that was email.
13:53So you can't-- you've
got to keep trying.
13:57You got to keep experimenting.
13:58You have to keep current.
13:59KARIM R. LAKHANI: Adi,
such a good point.
14:02And I have to tell you, around
COVID I had this experience
14:06with my mother, who lives
in Toronto, and my in-laws,
14:09who also live in Toronto, when
the things sort of eased up
14:13a bit and in November of 2022
when Thanksgiving was on we
14:21were able to fly them back to
have a reunion with them and--
14:26in 2020, sorry, November
2020, when things eased up
14:30a bit with the vaccines
and so on and so forth,
14:32we were able to bring them over.
14:37And [INAUDIBLE] dates,
but around that time,
14:39one Thanksgiving in COVID.
14:41And if you recall, traveling,
even just crossing the border
14:45from Canada to the
US was very difficult
14:48because you had QR
codes and apps galore.
14:52Canada needed all
this stuff to exit,
14:54US needed these stuff to
enter, and the same thing
14:58for entry as well.
15:00And I looked at, sadly,
how helpless my in-laws
15:05my parents were with
these technologies.
15:07They were just lost.
15:09And my wife and my daughter and
I had to spend a ton of time
15:12with them, holding their hands
to go through these things.
15:15Now, of course,
the UX was terrible
15:16and all that kind of stuff.
15:17But we figured out how to do it
ourselves, but they were stuck.
15:21They couldn't adjust.
15:22And then as I reflected
on that experience,
15:25I said, oh, this is what's
happening to most executives.
15:28This is what's happening
to most companies.
15:30It's like they're the senior
citizens, the elderly who
15:35have resisted the
technology, have not really
15:37embraced it, and now have no
choice but to deal with it
15:40and are frozen and
need a ton of help.
15:43And that's what the thing
that we have to get over as we
15:47think about this.
15:48ADI IGNATIUS:
Yeah, that's great.
15:49I think we can
throw out our QCADs.
15:50But otherwise--
15:51KARIM R. LAKHANI: Yes.
15:52I remember, and I
have a QCAD still.
15:55Old jokes for those
in publishing.
15:56ADI IGNATIUS: Yeah, I know.
15:57That's definitely--
this dates us.
16:00So let's talk about
generative AI.
16:01So people who, if that sounds
jargon-- oh, by the way,
16:04you referred to
earlier to Tsedal,
16:06and I just want to fill out
of that's Tsedal Neeley,
16:08who is another--
16:09KARIM R. LAKHANI: Professor
Tsedal Neeley, Senior Associate
16:11Dean of the Harvard Business
School, my good friend,
16:13and author of another great
book called The Digital Mindset.
16:16ADI IGNATIUS: Absolutely.
16:17So all right, generative AI, for
people who aren't conversant,
16:20and probably most of you
are now, but this is large
16:26language model learning,
predictive, ChatGPT, Bard,
16:31Bing, all these things.
16:32And I like to say that
there were three waves.
16:36The first was we played
with this technology
16:39when it came out in whatever
it was, November of last year.
16:42And we tried to break it.
16:43ChatGPT, are you
in love with me?
16:46And now we're trying to
figure out how to use it,
16:49and the use cases are happening.
16:51So first question for you,
where are you in the hype cycle?
16:54Because technologies
come and go.
16:56And there was a
sense with this one
16:58that everyone--
everyone-- basically
17:00said, OK, this is different.
17:02This is important.
17:03This is transformational
in ways that
17:05few technological
innovations are.
17:06But where are you on that?
17:08KARIM R. LAKHANI: I'm
on the, holy crap,
17:10this is transformational.
17:14Yeah.
17:15So look, the way
I think about this
17:18is it's actually worth to pause
and look at history a bit.
17:23And since I studied
technology in business,
17:25something transformational
happened 30 years ago,
17:27approximately, as well,
which was the browser.
17:30The browser got invented.
17:32And if you think
about the browser,
17:34there was 30 years
of the internet.
17:36The browser gets invented,
and people were like, oh,
17:40my goodness, look at this.
17:41And I remember-- I can
still see as clear as day
17:45when I first
encountered the browser.
17:48And I was working
at General Electric.
17:50I was at a conference
for radiology,
17:52and one of my
clients, a radiologist
17:56at Saint Paul's
Hospital in Vancouver,
17:57showed me the browser
and the thing he showed
17:59me was the Oxford coffee pot.
18:01I'm like, interesting.
18:03All of a sudden, the
Oxford coffee pot
18:05has global distribution.
18:07Anybody that has a web browser
and an internet connection
18:10can use it.
18:11So there's 30 years of
internet in the basement
18:13in the bowels of companies.
18:15We didn't understand it.
18:16We saw it was coming,
it was coming.
18:17It was USENET, there was
Gopher, there was Telnet,
18:20there was FTP, all
these kinds of things.
18:23The browser showed what
the world would look like.
18:26And the initial applications
were cute applications.
18:29And people were,
like, this is nothing.
18:31This is whatever.
18:32But fundamentally,
from an economics point
18:34of view, what the
browser did is that it
18:36lowered the cost of information
transmission dramatically.
18:40And then the last
30 years, we've
18:43been living through the
buildout of the internet
18:45and waves and waves of
the internet changing
18:49more and more and more
industries over and over again.
18:52We've all living through that.
18:53Just imagine this right now.
18:56We are broadcasting
live to, I don't know,
18:58thousands of people,
and more people
19:00will be looking at this
broadcast at relatively zero
19:05marginal cost to us to do this.
19:08It seems unbelievable
compared to 1993,
19:11where you needed a massive
TV studio, massive broadcast
19:14studio, satellite
dishes to be able to do
19:16what we're doing right now.
19:18And so the cost of information
transmission went to zero.
19:22And then new companies formed,
Google, Amazon, Facebook,
19:26you name it.
19:26E-commerce got invented
and so on and so forth.
19:30And so that is the world
that we are coming out of.
19:34The internet era is
we're coming out of.
19:36And the same thing has
happened with generative AI.
19:38There's been 20 years
of AI being deployed
19:42at scale inside of many tech
companies, the ones we use
19:45in our examples in our book.
19:47And that was in the basement, so
Netflix movie recommendations,
19:52your Google search results,
your Amazon recommendations,
19:56your Spotify music
results, your car access.
19:59All of that was being--
20:00your Waze access,
your directions,
20:02all that was being
empowered by AI tools, even
20:07your spam killers.
20:08Remember how bad spam
used to be for a while,
20:11and then overnight it went away?
20:12Because people deployed machine
learning systems to those
20:15things, early machine learning
systems to those things.
20:17Now--
20:19ADI IGNATIUS: Go ahead.
20:20KARIM R. LAKHANI: Now, how do
we think about generative AI?
20:22So my view is, generative
AI is a drop in the power
20:27in the cost of cognition,
in how we think.
20:30So if the internet was one of
cost of information dropping
20:35to zero, my sense is that
the cost of cognition,
20:39how we think, who we think
with, is dropping to zero
20:43or lowering
significantly with this.
20:45And that has significant
ramifications for this.
20:50And I have to tell you, I
had to do a major pivot even
20:55in my research side on
what to do with this.
20:58I was doing a lot of
stuff around AI adoption
21:00and so on and so forth,
a lot of research,
21:02a lot of nerdy research that
only, like, three people ever
21:04read.
21:05But we have gone--
21:08my whole Institute and my
labs have gone big time
21:11into figuring out what this
means for knowledge workers,
21:13for managers with generative AI.
21:15ADI IGNATIUS: Yeah,
well, let's talk
21:16about that because I think,
again, for our viewers,
21:19there are probably some of you
are well conversant or using
21:22it, whether it's for fun
or to try to figure out
21:24work applications.
21:25There are products
available that
21:28are using this that
rolled out pretty quickly.
21:31How to think-- is there a way
to think about this generically?
21:34What is, for a generic company,
if there is such a thing,
21:40how should they think
about using generative AI?
21:44I mean, we've seen the
amazing things that can do.
21:46We've seen the hallucinations
that they can create,
21:50that confidently provide errors.
21:56So how does a
business think about,
21:58is there a generative
AI application that
22:01is significant for my company
and how do I figure that out?
22:03KARIM R. LAKHANI: Yeah.
22:04So first of all, I think we're
at the super early stages
22:06of this hype, of this cycle.
22:08And if you think about
it, the first web browser
22:10was Mosaic, and then Netscape
and then Explorer and Mozilla
22:15and so on and so
forth came about.
22:16So I think we
should just think--
22:18and then all the
applications on top.
22:19So I think we are
at the early stages.
22:23The rate of innovation and
the rate of improvement
22:26is increasing rapidly,
and it keeps increasing.
22:30And the rate of
application development
22:31is also increasing rapidly.
22:33So the first thing
I would say is
22:34that the places where you can
apply it is in many ways like,
22:38well, where do you
apply thinking?
22:39Well, where else
could you apply this,
22:41with all the caveats about
hallucination and bias
22:44and so forth.
22:45So the first thing
I would say is
22:46that I think, if you
step back and say,
22:50what should leaders do,
what should managers
22:52do, what should executives
do around this thing,
22:54one is to start
thinking about and start
22:56practice in their own sandboxes
what the use cases may be.
23:01We're seeing tremendous
use cases, for example,
23:04just in sort of content
generation, like, our work.
23:06Us as knowledge producers,
that's changing rapidly.
23:11Now I use ChatGPT as an
amazing research associate,
23:16thought partner, copy
editor, idea generator.
23:21I can tell you one thing.
23:22I was in Asia.
23:24My wife was with
me on my trip and I
23:26wanted to actually have
some time for break as well.
23:30So I went to ChatGPT
and I said, this is me.
23:33This is my wife.
23:34Here's the kind of
vacations we like.
23:36Can you please give us
ideas of a place that
23:39would be about three hours from
Singapore that we could go to?
23:42And I prefer beach,
blah, blah, blah, blah.
23:44Boom.
23:45In microseconds, I got
many recommendations.
23:48And then through a
conversational setup,
23:52I found the place that
we wanted to go to.
23:53It was a hidden place in the
South China Sea off Indonesia,
23:57and it was incredible.
23:59It was incredible.
24:00And that I would
not have discovered,
24:02even with my travel agent.
24:04So just even in
that activity, just
24:05imagine what we can now
start to do with this.
24:08And so what I would say is
that the thing that managers
24:11and leaders need to do is,
step 1, start using it.
24:16I think the bans on
ChatGPT and on these things
24:19are misguided in many companies.
24:21It's already on my phone.
24:23It's already people--
there's 100 million users.
24:26It's already there.
24:28So I think executives
and IT departments
24:30and legal departments
are fooling themselves.
24:32They don't think their
workers are already
24:34not using these tools.
24:35And instead of pushing
against it and saying,
24:38no, you need to embrace
it and run boot camps,
24:41run use case
analysis, figure out
24:44where it's hallucinating
in your use cases
24:46and figure out
where it's actually
24:47going to be very helpful.
24:48And what I say to people, for
managers, leaders, and workers,
24:52is AI is not going to replace
humans, but humans with AI
24:56are going to replace
humans without AI.
24:58And this is definitely the
case for generative AI.
25:00And so the first step is
begin, start experimentation,
25:04create the sandboxes,
run internal boot camps.
25:06And don't just run boot
camps for technology workers.
25:09Run boot camps for everybody.
25:10Give them access to tools.
25:12Figure out what use cases
they develop, and then
25:14use that as a basis
to rank and stack them
25:17and put them into play.
25:18ADI IGNATIUS: Yeah,
I agree with that.
25:20We have to think about
that as a publisher.
25:21There are some
publishers who say,
25:23we will not take articles,
papers where generative AI has
25:32been involved.
25:33That, similarly to me,
doesn't make sense.
25:34It's like saying, don't use--
25:35KARIM R. LAKHANI:
How will they know?
25:36How will they know?
25:37ADI IGNATIUS: How
will they know?
25:38And then it would be like
saying, don't use Google.
25:39It's a tool.
25:40What we're saying, though,
is that the responsibility
25:43more than ever is on the person
with the byline on this piece.
25:48That was true-- you didn't
want to just use Google search
25:50results or just use
Wikipedia results.
25:52You need to verify and
do a little bit more
25:55than that, now more than ever
because you may be relying on--
25:58KARIM R. LAKHANI:
No, absolutely.
25:59And as scholars,
we publish, again,
26:01these nerdy papers that
very few people read,
26:03and we're in the same
crisis because, well,
26:07if I use an RA to
come up with ideas,
26:09do I have to acknowledge the RA?
26:11Is the RA the co-author?
26:12If I use a copy
editor, I typically
26:14don't acknowledge a copy
editor for my article,
26:16but they're super helpful.
26:18Should that be-- attribution
becomes interesting.
26:22There's so many important
questions at play,
26:24just as writers and producers.
26:26But my advice to everybody is--
26:28and the best place to
learn, Adi, is YouTube.
26:32YouTube has, oh my God, so many
tutorials and so many domains.
26:37And very soon, you'll
be down the rabbit hole.
26:39Is learn and adopt and
practice and practice and see.
26:45ADI IGNATIUS: Yeah.
26:46And I think I think there's
a trap that sometimes people
26:48feel like, if they don't
jump on the wave immediately,
26:51somehow it's too late.
26:53KARIM R. LAKHANI:
No, no, gosh, no.
26:54ADI IGNATIUS: And I think it's
really important-- it's early.
26:57KARIM R. LAKHANI: Oh my gosh.
26:57ADI IGNATIUS: Karim, if you're
right that this is truly
26:59transformative, it's early.
27:01If you feel like,
wow, everybody's
27:02moving faster than
I am, catch up,
27:04whether it's YouTube or
just doing some reading
27:06and figure out how to applies.
27:07So let me go to some
questions from the audience
27:09because there are
a lot coming in.
27:10And so this is from [? Vena ?]
from somewhere in the US.
27:15And it's really
[? Vena ?] is saying
27:17that AI, machine learning,
this is somebody's code.
27:20It comes with biases and
assumptions built in.
27:24It's a topic we
think about a lot.
27:26But what's your view?
27:27How can the industry
ensure that there
27:30isn't a monopoly on how we
think and how we're biased
27:35and the assumptions
that we make?
27:36KARIM R. LAKHANI: Yeah.
27:37So look, I as an individual, I'm
part of Mozilla, mozilla.com.
27:44We made the Firefox browser
owned by an open source
27:48foundation, Mozilla Foundation.
27:51If you love Fire--
if you haven't
27:52used Firefox in a while,
go back and use it again.
27:54But we just set up
Mozilla.ai, and the idea
27:57is that we want to create open
source large language models as
28:00well and create the
tooling that enables
28:04many people around the world
to also have large language
28:07models suited for them.
28:08And our view is that we can
build tools to, A, detect bias
28:13and to fix bias and to
fix all the craziness
28:16that these large
language models can do.
28:18So I'm actively working and
trying to create and support
28:21organizations that do that.
28:23I would say that the first
thing that we need to think,
28:27though, is to step back and
say, the world is biased.
28:33We had bias before there was AI.
28:35AI is just amplifying it
and making it apparent.
28:38The world is biased.
28:39You look at the unbelievably
bad treatment African-Americans
28:44receive in our health care
system and the financial system
28:46and so on and so
forth in the US,
28:48or if you go to
some other country,
28:50there's always been
discrimination without AI.
28:53AI is helping to amplify it.
28:54And so the response-- the
ethical responsibility
28:56for us as leaders has to be
that we have to understand what
29:00is biased today in our systems.
29:03And then let's translate to AI.
29:05Well, certainly, how
representative is your data?
29:08How representative
is your training?
29:10How representative
is your labeling?
29:11Those are essential,
essential questions
29:14that need to be
now part of, again,
29:16the executive conversation.
29:17That's why the learning
mandate doesn't stop,
29:19because you have to understand
how these machine learning
29:22systems are built for you to
understand what the biases are
29:25and how you might get sued or
be put in jail, for God sakes,
29:28if you don't follow
through on these things.
29:30And so this is super important.
29:32But I want us to
also be aware that we
29:35need to think counterfactually.
29:36There's always this
bias in the world.
29:38And now let's imagine
a world with AI.
29:41And is it going to take the
biased world and amplify it?
29:44Or can we correct for it?
29:46Can we recognize it for it?
29:47So that's going to
be very important.
29:48ADI IGNATIUS: Yeah.
29:49And it depends,
maybe, on ultimately
29:50if you're a techno optimist
or a techno pessimist.
29:53KARIM R. LAKHANI: Yes.
29:54I tend to be on
the optimist side.
29:56ADI IGNATIUS: Yeah.
29:57So there are a million questions
coming in on this topic,
30:00and we're not going to be able
to get to them all or even
30:03close to all.
30:04But this is sort of a different
question that's come in,
30:06but I think it's
kind of interesting.
30:07This is from Janelle
in Washington, DC.
30:09So when we're talking
about dealing with waves
30:14of technology and
changing and adapting,
30:16so the question
from Janelle is--
30:20we've been talking about
how your employees can
30:22learn and adapt.
30:23But how do you help
the customer learn?
30:25Because sometimes
that's the learning arc
30:28that you need to accelerate
for your business to go
30:30to where it wants [INAUDIBLE].
30:32Do you have thoughts
on how to [INAUDIBLE]??
30:34KARIM R. LAKHANI: That
is a great question.
30:35I think customers tend
to be ahead, I think.
30:37I think what happens,
what's interesting--
30:39I was in sales and marketing for
four years at General Electric.
30:42And what would happen, what
would be so interesting
30:45is, because my customers
knew what I had
30:50and what I didn't have
and what we're good at
30:52and what we're not
good at, they wouldn't
30:54talk to me about things
that we were not good at
30:56or we weren't exploring.
30:57So I never got that message
until much, much later.
31:01I discovered, like, oh,
you're interested in this?
31:03Oh, we've got some
nascent product, whatever.
31:05But they said, no, we knew GE
was not going to be good here,
31:08so we didn't talk to you.
31:10So I think you'd be surprised,
especially today's customers
31:14because, again, as I mentioned,
all of them are living,
31:17including yourselves,
Janelle, are
31:19living in a digital age
with our smartphones
31:21and our capabilities
and so forth.
31:23So you'd be surprised
at how fast they adapt.
31:26And in many situations,
with other companies
31:29they're already further
down the pike than with you.
31:31And so we get the wrong
signals from our sales teams,
31:34from our marketing teams,
even from our focus groups,
31:37because we actually don't
observe customers in situ
31:39and see what's going on.
31:42And again, you think about
the median user of Facebook,
31:48I think they're, like, 50
or something right now.
31:51So I think adoption is no
longer as big a deal that I
31:59think we think it is.
32:00ADI IGNATIUS: My generation
ruined Myspace and now
32:02we're going to ruin [INAUDIBLE].
32:03KARIM R. LAKHANI: I know.
32:03Look at that.
32:04Now we're aging
out Facebook, too.
32:06ADI IGNATIUS: We'll
get TikTok next.
32:08So last question.
32:09And again, I wish
we had more time.
32:11But maybe the new
wave of how people
32:14are thinking about
generative AI is,
32:16and at first with
an explanation,
32:18it's really like that technology
that predicts the next word--
32:21KARIM R. LAKHANI: Yes, totally.
32:22ADI IGNATIUS: --sort
of on steroids.
32:24But then it sort of evolved
into this almost feels
32:28like a sentient--
32:30the machine is developing this
kind of emotional intelligence
32:34or that we may be
on the path to that.
32:36Is that-- what's your view?
32:37Is that a pure illusion or are
we heading toward something
32:41that will at least feel
like an intelligence
32:44and an emotional
intelligence [INAUDIBLE]??
32:46KARIM R. LAKHANI:
What a good question.
32:48The first thing the first
thing I always say is,
32:51be kind to your robots.
32:53So always say
please and thank you
32:55when you're using
ChatGPT or Bard.
32:57I do that as a principle.
32:58I tell everybody, be
kind to your robots,
33:00because if the sentience
moment shows up,
33:08all the data will be there.
33:09All the history of our
records with these systems
33:11will be there.
33:12And you don't want them to get
pissed off because, hey, Karim
33:14was a bad actor for us.
33:16So always be-- like,
I'm an ardent atheist,
33:21but I still say inshallah.
33:23ADI IGNATIUS: Just in case.
33:24KARIM R. LAKHANI: Who knows?
33:25Just in case.
33:26Hedge your bets.
33:27Hedge your bets.
33:29So like I say, inshallah,
we should always
33:31say please and thank
you to your robots.
33:34That's the first thing.
33:35The second thing is, right
now the human-like responses
33:41are a statistical illusion.
33:42They absolutely are.
33:43They've just been
well trained by humans
33:45to respond like humans.
33:47And they've used all of our
texts and all of our videos
33:50to be human like in many ways.
33:52But in the end, it's a
statistical or computational
33:57illusion.
33:58But I can tell you, I got
a little bit of a wake
34:01up call on this.
34:02I felt like this stuff,
like, the strong AI
34:04stuff that has been talked
about this-- all this stuff
34:07is what we call weak AI.
34:08The strong AI stuff is decades
away, multi decades away.
34:14But in conversations
with leaders
34:16at Harvard at the
Kempner Institute,
34:18which is the new Institute
for Natural Intelligence
34:20and Artificial Intelligence, so
marrying biology with AI and AI
34:23with biology, so two amazing
scholars, amazing world
34:27leaders both in neuroscience
and machine learning, said,
34:30hey guys-- this is pre-ChatGPT
as well, by the way.
34:33Said, hey, guys, what
do you guys think?
34:35How far away is this
strong AI world?
34:38They said, 20 years.
34:40And I was like, whoa.
34:41I'm not ready for that.
34:45But the world
experts, people that
34:46know better than I do on this
thing, are saying 20 years.
34:49It might even be faster.
34:50The thing that's
interesting to me,
34:52Adi, is we may not even
know when it has sentience,
34:57because it's like we
assume human-like forms
35:00on alien intelligence,
on intelligence.
35:03But if you read a lot of
science fiction, like I do,
35:09maybe alien life is going to
be carbon-based, but maybe not.
35:13Maybe they'll have a
different metabolism, maybe
35:16different neural systems.
35:17And you need to
be ready for that.
35:19So we may not even know
it, that's the thing.
35:21And I think-- there was a paper
that was at Harvard recently
35:25called "First Contact."
35:26Some machine learning expert
at Microsoft wrote this paper.
35:30And I was like, wow,
that's pretty radical.
35:32ADI IGNATIUS: Yeah.
35:33This is fabulous.
35:34We're going to have to
get you back on the show
35:35because there's a lot
more to talk about.
35:37KARIM R. LAKHANI:
Yeah [INAUDIBLE]..
35:38ADI IGNATIUS: We didn't even get
into the Congressional hearings
35:40on aliens.
35:41KARIM R. LAKHANI: Oh, gosh yes.
35:42ADI IGNATIUS: We're just a
half step away from that.
35:44So this has been Karim Lakhani,
Harvard Business School
35:48Professor.
35:49Karim, thank you very much
for being on the show.
35:51KARIM R. LAKHANI: Great
to be here with you, Adi.
35:52ADI IGNATIUS: OK.
35:53See you soon.
35:54All right.
35:54So I want to thank you
all for joining us today.
35:56Just a reminder, you can
watch previous episodes
35:59of the show on YouTube or right
here on LinkedIn and Facebook.
36:02Now, be sure to join us next
week on Wednesday, August
36:059 at 12:00 noon Eastern
time, when my guest will
36:09be Andrew Liveris he's the
former CEO of Dow Chemical, who
36:12in that job got
credit for pushing
36:14an ambitious sustainability
agenda at the company.
36:18He's also the author
of a new book,
36:19Leading Through
Disruption-- a Changemakers
36:22Guide to 21st
Century Leadership,
36:24and we'll be talking about
how inspired leaders can best
36:26adapt to the challenges that
keep getting thrown their way.
36:30Again, if you're an HBR
subscriber watching this,
36:32you can head to
hbr.org/newsletters to sign up
36:36for The New World
of Work newsletter,
36:37where I offer an inside look at
each of these interviews each
36:40week and talk about some of the
ideas that came out of them.
36:42And again, if you like
content like this,
36:44Why not subscribe to our
magazine and website.
36:46The address is
hbr.org/subscriptions.
36:50And finally, we want
to thank our friends
36:51at KPMG, who are our sponsors,
for this season of The New
36:55World of Work.
36:56And I want to thank all of
you for tuning in today.
36:58I'm Adi Ignatius, and this
is The New World of Work.
37:01[MUSIC PLAYING]
🎥 Related Videos
Hip Pain RELIEF! Stretches And Exercises For Lateral Hip Pain
Jenkins-Day-1|| Introduction to Continuous Integration and Continuous Deployment
How to Build the Life You Want: Timeless Wisdom for More Happiness & Purpose
Discovery 臺灣部落寶藏 第1季 信仰祭儀
Lessons Learned at Coursera – Expert Roundtable with Julia Stiglitz
Reality Transurfing Chapter 14 "Forward to the Past" by Vadim Zeland
🔥 Recently Summarized Examples
4 Steps to Master Any Complex Skill (quickly)
40 Years of Fitness Experience in Less Than 11 Minutes.
Gun Controlling Media Makes FATAL Mistake... They Have Tied Their Fate To Biden's & Gun Rights Win
GET READY! Palantir Is Officially The Next Nvidia.
Abundant Thinking: The Hidden Key to Get Everything You Want (Audiobook)