Karim Lakhani is a professor at Harvard Business School who specializes in workplace technology and particularly AI. He’s done pioneering work in identifying how digital transformation has remade the world of business, and he’s the co-author of the 2020 book Competing in the Age of AI. Customers will expect AI-enhanced experiences with companies, he says, so business leaders must experiment, create sandboxes, run internal bootcamps, and develop AI use cases not just for technology workers, but for all employees. Change and change management are skills that are no longer optional for modern organizations.
Just as the internet has drastically lowered the cost of information transmission, AI will lower the cost of cognition. That’s according to Harvard Business School professor Karim Lakhani, who has been studying AI and machine learning in the workplace for years. As the public comes to expect companies that deliver seamless, AI-enhanced experiences and transactions, leaders need to embrace the technology, learn to harness its potential, and develop use cases for their businesses. “The places where you can apply it?” he says. “Well, where do you apply thinking?”
For this episode of our video series “The New World of Work”, HBR editor in chief Adi Ignatius sat down with Lakhani, author of Competing in the Age of AI: Strategy and Leadership When Algorithms and Networks Run the World, to discuss:
• How executives and regular employees can (and must) develop a digital mindset
• Change management as a critical skill that must be in the DNA of any successful organization
• The shapes AI may take in the near and far future
READ MORE https://bit.ly/4moreeee
ADI IGNATIUS:
Karim, welcome to the show.
Karim Lakhani:
So glad to be here with you today, Adi.
ADI IGNATIUS:
You co-wrote a piece for us a few years ago, and it’s reflected in your book, where you say machine learning has basically changed the very rules of business. That’s a big statement. What do you mean by that?
Karim Lakhani:
The book really was a partnership between Marco Iansiti
and Amy Bernstein, one of the editors at HBR. And what
Marco and I noticed in about a decade's worth of research
and spending time with companies, both writing cases as
advisors, as consultants, and so forth, was that the nature of
the corporation (which really was established as the
modern American corporation, which became the blueprint
for the modern international corporation, established in
the 1920s and 30s) was changing foundationally because of
technologies like AI, like machine learning.
What we observed was that the entire business architecture
in many of these AI-first companies at that time, in terms of
business model, how you create value, how you capture
value, and your operating model, how you deliver value
how you achieve scope, the number of customers you serve,
the number of products you have, scale, the number of
customers you serve, and learning these fundamental parts
of a business architecture, were being rewired because of
machine learning, AI, and digital technologies.
If you just reflect for a bit on your experience using Google
for example, much of your Google experience is fully
automated, from the ads you see to the search you do to, if
you're using Gmail, how you interact with them. It's not
people that do those activities, it's the algorithms that make
that happen. Similarly with allthe large e-commerce
platforms like an Amazon or Alibaba or Netflix. But these
companies work in a fundamentally different way than a
company like General Electric, where I grew up right out of
college in my first job.
These companies, the machines and the algorithms are at
the center. The work is automated. The humans are actually
designing the algorithms and testing them and checking
them, making sure they're working within bounds, but the
actual transactions and activities are being mediated
through the machines.
ADI IGNATIUS:
Uber car or your Ola car doesn't show up in three minutes."
You want this magical taxi experience. You go on your app
you press the thing, boom, it shows up. And ifit's going to
be five or seven minutes, you kind of get mad. I'm reflecting
on when I first moved to Boston in 1997, and it would take
me a week to book a taxi in Boston. Now, we get mad.
Similarly, if there's a transaction dispute on Amazon or
Uber, it's automatically solved, done. But the same people
the same executives, in their own companies are
completely satisfied if a customer service interaction can
take two weeks, if onboarding a new vendor takes six
months. We're living in this disconnected world where most
people. most consumers, are living in an AI-first world in
their experiences with many of these platforms. And then
they encounter our companies and our organizations, and
they're like, "What is this?"
My sense is that this is inevitable. This transition is really
inevitable. And for the folks that are behind, the good news
is that the cost to make the transition keeps getting lower
and lower. The playbook for this is now well-known. And
finally, the real challenge is not a technological challenge.I
would say that's like a 30% challenge. The real challenge is
70%, which is an organizational challenge. My great
colleague Tsedal Neeley talks about the digital mindset.
Every executive, every worker needs to have a digital
mindset, which means understanding how these
technologies work, but also understanding the deployment
of them and then the change processes you need to do in
terms of your organization to make use of them.
ADI IGNATIUS:
We're a relatively small publisher compared to some of the
giants out there. But when people come to our site, and
they're searching for articles by Karim Lakhani, they're
used to a Google sort of search experience
Karim Lakhani:
Exactly.
READ MORE https://bit.ly/4moreeee
ADI IGNATIUS:
When they want to buy a product from us, they're used to
an Amazon, and anything short of that, it's like your Uber
example. There's a frustration and expectation. So we have
to. find ways to lift our game without the resources, whether
it's through partnerships or other things, because it's table
stakes. People just expect the best experience in every
experience they have.
Karim Lakhani:
One hundred percent, and if I look at my teenage daughter,
she has no patience for old line companies. She just gets
mad and is like, *What is this?"
ADI IGNATIUS:
The next big wave is generative AI. But that won't be the
last wave, and quantum will hit us at some point, and
things we can't even anticipate will hit us. How do you
prepare for that? How do you create a culture or mindset or
organization that knows there will be unexpected waves of
technology, we'll have to figure out if they're relevant to us
or not, and if they are, we need to adapt quickly? Is there a
general way to think about that?
Karim Lakhani:
I think there are two imperatives for most executives, for
most managers, for most leaders. One is a learning
imperative. This is, again, Tsedal [Neeley]'s work on digital
mindset, and my work. There's lots of learning you need to
do, and the learning has to be continuous. The idea is not
that we want you to become AI engineers or data scientists
or get a Ph.D., but as executives, this is now table stakes.
The way I think about this at the MBA program for us, is
that people come to the Harvard MBA program and we have
a first year required curriculum. There are 10 courses and
one of them is accounting. I can tell you, accounting is a
very important profession, but most MBAs that join HBS
don't want to be accountants. But they need to learn
accounting because that's the language of business. That's
the way in which you think about how value is kept track of
how expenses are tracked, and so on. Super important.
You don't take the accounting course to become
accountant, but you need to understand accounting so you
can be a good business person. Same thing now with digital
technologies and machine learning. You need to
understand the machine learning stuff and the AI stuff, not
because you're going to become an AI engineer or an AI
scientist, but because that is now going to be a critical table
stakes for you to understand how business works
There's a learning imperative and I don't think we can take
away the learning imperative anymore. I'm self-centered
about this. I'm self-interested. I'm in the learning
profession. So, I want to caveat that. But I want to just insist
that the learning journey does not stop and you have to
invest in your own personal learning, and Ithink
companies need to invest in the learning for their own
employees as well. It's a two-phased conversation.
Companies have to embrace this and so do individuals
But the second bit is equally important, which is
completely underrated: change and change management
become skills for managers and for leaders and for
executives. How you change, how you continuously
change, how you build a DNA for changing becomes very
important.
I was in Asia a month ago and had a chance to spend some
time with Mickey Mikitani at Rakuten. He has thought
about change as a core competency for all workers and all
employees. Right now, most change programs are viewed
with skepticism, flavor of the month, blah, blah, blah.
People resist.
The best companies will be the ones that can understand
how change becomes a skill. If you think about change as a
skill, what does that mean? Skills require acquisition of the
skills. You've got to invest in learning. What does it mean to
change? It requires practice. You've got to keep changing as
well. And it requires adjustment. Those elements I think
will become a key part of the ways in which leaders need to
adapt to this world.
ADI IGNATIUS:
It could be generational but it should not be generational. I
always think that when you come of age, when you join the
workforce, there's a certain suite of technology that you
grew up with and you're comfortable with and you are part
of figuring out how to use it. Subsequently, a lot of us hit a
point where a new one seems stupid. For my father who's
still alive at 102, that was email. Right?
Karim Lakhani:
Yes.
ADI IGNATIUS:
So you've got to keep trying. You've got to keep
experimenting. You have to keep current.
Karim Lakhani:
Around Covid I had this experience with my mother who
lived in Toronto and my in-laws were also living in
Toronto. When things eased up a bit in November of 2020
when Thanksgiving was on, we were able to fly them back
to have a reunion with them. When things eased up a bit
with the vaccines and so on and so forth, we were able to
bring them over. And if you recall, traveling, even just
crossing the border from Canada to the US, was very
difficult because you had QR codes and apps galore. Canada
needed all this stuff to exit, US needed it to enter.
And I looked at how, sadly, helpless my in-laws and my
parents were with these technologies. They were just lost
and my wife and my daughter and I had to spend a ton of
time with them, holding their hands, to go through these
things. Of course the UX was terrible, but we figured out
how to do it ourselves. But they were stuck. They couldn't
adjust.
As I reflected on that experience, I said, "Oh, this is what's
happening to most executives. This is what's happening to
most companies." They're like the senior citizens, the
elderly, who have resisted the technology, have not really
embraced it, and now have no choice but to deal with it and
are frozen and need a ton of help. That's the thing we have
to get over as we think about this.
ADI IGNATIUS:
OK, generative AI. I like to say that there were three waves.
The first was that we played with this technology when it
came out. Then we tried to break it by asking, "ChatGPT.
Are you in love with me?"
Karim Lakhani:
Yes.
ADI IGNATIUS:
And now we're trying to figure out how to use it. Where are
you in the hype cycle?
Karim Lakhani:
I'm like, "Holy crap! This is transformational." The way 1
think about this is that it's actually worth it to pause and
look at history. Since I studied technology and business
something transformational happened 30 years ago
approximately as well, which was the [web] browser got
invented. If you think about the browser, there were 30
years of the internet, then the browser gets invented, and
people were like, "Oh my goodness. Look at this."
I remember I could still see as clear as day when I first
clicked on the browser, and I was working in General
Electric, I was at a conference for radiology, and one of my
clients, a radiologist at St. Paul's Hospital in Vancouver,
showed me the browser, and the thing he showed me was
the Oxford Coffee Pot. I'm like, ""Huh. Interesting." All of a
sudden, the Oxford Coffee Pot has global distribution
Anybody that has a web browser and internet connection
can use it.
There's 30 years of internet in the basement, in the bowels
of companies. We didn't understand it. We saw it was
coming, it was coming. It was Usenet, there was Gopher,
there was Telnet, there was FTP, all these kinds of things.
Then the browser showed what the world would look like.
And the initial applications were cute applications and
people were like, "Ah. This is nothing. This is whatever." But
fundamentally, from an economics point of view, what the
browser did is that it lowered the cost of information
transmission dramatically.
Then in the last 30 years. we've been living through the
build out of the internet, and waves and waves of the
internet changing more and more industries, over and over
again. We've all living through that,
We are broadcasting live to, I don't know, thousands of
people, and then more people will be looking at this
broadcast, at relatively zero marginal cost to us. This seems
unbelievable compared to 1993 where you needed a massive
TV studio, massive broadcast studio, satellite dishes to be
able to do what we're doing right now. The cost of
information transmission went to zero, and then new
companies formed, Google, Amazon, Facebook, you name
it, e-commerce got invented. That is the world we are
coming out of.
The same thing's happened with generative AI. There's
been 20 years of AI being deployed at scale inside of many
tech companies. That was in the basement. Netflix movie
recommendations, your Google search results, your
Amazon recommendations, your Spotify music results,
your car access, your Waze access, your directions. All that
was being empowered by AI tools. Even your spam killers.
Remember how bad spam used to be for a while, and then
overnight it went away? Because people deployed machine
learning systems.
How do we think about generative AI? My view is
generative AI is a drop in the cost of cognition and how we
think. If the internet was the cost of information dropping
to zero, my sense is that the cost of cognition, how we think.
who we think with, is dropping to zero, or lowering
significantly.
That has significant ramifications. I had to do a major pivot
even in my research on what to do with this. I was doing a
lot of stuffaround AI adoption. A lot of research, a lot of
nerdy research that only three people will read, but my
whole institute and my labs have gone big time into
figuring out what this means for knowledge workers, for
managers with generative AI.
ADI IGNATIUS:
Let's talk about that. There are products available that are
using this, that rolled out pretty quickly. Is there a way to
think about this generically? For a generic company, if there
is such a thing, how should they think about using
generative AI?
Karim Lakhani:
First of all, we're in the super early stages of this hype, of
this cycle. If you think about it the first web browser was
Mosaic, and then Netscape and Explorer and Mozilla came
along. Then all the applications on top. We are at the early
stages. The rate of innovation and the rate of improvement
is increasing rapidly, and it keeps increasing. The rate of
application development is also increasing rapidly.
The places where you can apply it is--well, where do you
apply thinking? Where else could you apply this, right?
With all the caveats about hallucination and bias and so
forth, if you step back and say, what should leaders do?
What should managers do, what should executives do
around this thing? One is to start thinking about and start
practice in their own sandboxes what the use cases may be.
We're seeing tremendous use cases, for example, just in
content generation. Our work, as knowledge producers,
that's changing rapidly. I use ChatGPT as an amazing
research associate, thought partner, copy editor, idea
generator.
I was in Asia, my wife was with me on my trip, and I wanted
to actually have some time for a break. I went to ChatGPT
and I said, "This is me, this is my wife. Here's the kind of
vacations we like. Can you please give us ideas of a place
that would be about three hours from Singapore that we ga
to, and I prefer a beach, etc." Boom. In microseconds, I got
many recommendations. And then through a
conversational setup, I found the place that we wanted to go
to, and it was a hidden place in the South China Sea off
Indonesia. And it was incredible. It was incredible. And that
I would not have discovered even with a travel agent. So
just even in that activity, just imagine what we can now
start to do with this.
The thing managers and leaders need to do is, step one,
start using it. The bans on ChatGPT and these things are
misguided in many companies. It's already on my phone
There are a hundred million users, it's already there. I think
executives and IT departments and legal departments are
fooling themselves if they don't think their workers are
already using these tools.
Instead of pushing against it and saying, "No," you need to
embrace it and run bootcamps, run use case analysis, figure
out where it's useful in your use cases and figure out where
it's actually going to be very helpful.
What I say to managers, leaders, and workers is: Al is not
going to replace humans, but humans with AI are going tO
replace humans without Al. This is definitely the case for
generative AI. The first step is to begin, start
experimentation, create the sandboxes, run internal
bootcamps, and don't just run bootcamps for technology
workers, run bootcamps for everybody. Give them access to
tools, figure out what use cases they develop, and then use
that as a basis to rank and stack them and put them into
play.
ADI IGNATIUS:
I agree with that. We have to think about that as a publisher.
There's some publishers who say we will not take articles or
papers where generative AI has been involved. That doesn't
make sense. It's like saying don't use Google. It's a tool.
What we're saying though is that the responsibility, more
than ever, is On the person with the byline on this piece,
That was true, you didn't want to just use Google search
results or just use Wikipedia results. You need to verify and
do a little bit more than that, now more than ever.
Karim Lakhani:
Absolutely. As scholars, we publish these nerdy papers that
very few people read, and we're in the same crisis. If I use a
research assistant to come up with ideas, do I have to
acknowledge them? Is the assistant a co-author? If I use a
copy editor, I typically don't acknowledge a copy editor for
my article, but they're super helpful. Attribution becomes
interesting. There's so many important questions at play
just as writers and producers
The best place to learn, Adi, is YouTube. YouTube has, oh
my God, so many tutorials in so many domains.
ADI IGNATIUS:
I think there's a trap that sometimes people feel like if they
don't jump on the wave immediately, somehow it's too late.
Karim Lakhani:
No, no. Gosh, no.
ADI IGNATIUS:
I think it's really early. If you're right that this is truly
transformative, it's early. If you feel like, "`Wow, everybody's
moving faster than I am," then catch up. Whether it's
YouTube or just doing some reading 1 5 and dlld iiguring figuring out how it
applies. Let Let me me go on th to co some questions from the audience
This is from Veena from somewhere in the US. AI is
somebody's code. It comes with biases and assumptions
built in. How can the industry ensure there isn't a
monopoly on how we think and how we're biased, and the
assumptions that we make?
Karim Lakhani:
I as an individual, I'm part of Mozilla. We made the Firefox
browser owned by an open source foundation, Mozilla
Foundation. If you haven't used Firefox in a while, go back
and use it again. We've just set up mozilla.ai, and the idea is
that we want to create open source large language models
and create the tooling that enables many people around the
world to have large language models suited for them. Our
view is that we can build tools to detect bias and fix bias,
and to fix all the craziness that these larger language
models can do. So, I'm actively working in trying to create
and support organizations that do that.
The first thing we need to do is step back and say the world
is biased. We had bias before there was AI. AI is just
amplifying it and making it apparent. The world is biased
You look at the unbelievably bad treatment African
Americans receive in our healthcare system and in our
financial system, and so on in the US. Or if you go to some
other country, there's always been discrimination without
AI. AI is helping to amplify it.
The ethical responsibility for us as leaders has to be that we
have to understand what is biased today in our systems
How representative is your data? How representative is
your training? How representative is your labeling? Those
are essential, essential questions that need to be part of the
executive conversation. That's where the learning mandate
doesn't stop, because you have to understand how these
machine learning systems are built for you, tor understand
what the biases are, and how you might get sued or be put
in jail, for God's sakes, if you don't follow through On these
things.
This is super important, but I want us to also be aware that
we need to think counter factually, there's always this bias
in the world. And now let's imagine a world with AI, and is
it going to take the bias world and amplify it or can we
correct for it? Can we recognize it for it? That's going to be
very important.
ADI IGNATIUS:
Yeah, this depends on if you're a techno-optimist or a
techno-pessimist.
READ MORE https://bit.ly/4moreeee
Karim Lakhani:
Yes, I tend to be on the optimist side
ADI IGNATIUS:
Here's a different question. This is from Janelle in
Washington, DC. We were talking about dealing with waves
of technology and changing and adapting. We've been
talking about how your employees can learn and adapt, but
how do you help the customer learn?
Karim Lakhani:
This is a great question. I think customers tend to be ahead
I was in sales and marketing for four years at General
Electric, and my customers knew what I had and what I
didn't have and what we were good at and what we were not
good at, they wouldn't talk to me about things that we were
not good at or we weren't exploring. I never got that
message until much later.
I discovered, "Oh, you're interested in this? Oh, we've got
some nascent product, whatever," but I said, ""No," because
we knew GE was not going to be good here, so we didn't talk
to you. I think you'd be surprised, especially today's
customers, because again, as I mentioned, all of them are
living in a digital age with our smartphones and our
capabilities. You'd be surprised at how fast they adapt, and
in many situations with other companies that are already
further down the pike than with you.
We get the wrong signals from our sales teams, from our
marketing teams, even from our focus groups, because we
actually don't observe customers in situ and see what's
going on. You think about the median user of Facebook,I
think they're like 50 or something right now. Adoption is no
longer as big a deal.
ADI IGNATIUS:
My generation ruined MySpace and now we're going to ruin
Facebook.
Karim Lakhani:
[ know. Look at that. Now, we're aging out of Facebook too
ADI IGNATIUS:
We'll get TikTok next. The last question: generative AI has
evolved until it almost feels sentient. Is this machine
developing emotional intelligence, or are we on the path to
that? Is that a pure illusion, or are we heading towards
something that will at least feel like an intelligence and an
emotional intelligence in these machines?
Karim Lakhani:
I always say to be kind to your robots, OK? Always say
please and thank you when you're using Chat GPT or Bard. I
do that as a principle. I tell everybody, be kind to your
robots because if the sentience moment shows up, all the
data will be there, all the history of our records with these
systems will be there. And you don't want them to get
pissed off because, "Hey, Karim was a bad actor for us." I'm
an ardent atheist, but I still say inshallah.
ADI IGNATIUS:
Just in case.
Karim Lakhani:
Who knows? Just in case, hedge your bets. Like I say
inshallah, we should always say please and thank you to
your robots. That's the first thing. The second thing is that
right now, the human-like responses are a statistical
illusion. They absolutely are. They've just been well-trained
by humans to respond to humans, and they've used all our
texts and all our videos to be human-like in many ways. But
in the end, it's a statistical or computational illusion
But I should tell you, I got a little bit of a wake-up call on
this. I felt like this stuff, like the strong AI stuff that has
been talked about, all this stuff is what we call weak AI. The
strong AI stuff is many decades away. But in conversations
with leaders at Harvard, at the Kempner Institute, which is
the new Institute for National Intelligence and Artificial
Intelligence. we talk about the marrying of biology with AI
and AI with biology. Two amazing scholars, I I asked (this is
pre-Chat GPT), "What do you guys think? How far away is
this strong Al world?"
They said 20 years. And I was like, "Whoa, I'm not ready for
that." But if the world experts, people that know better than
I do on this, are saying 20 years, it might even be faster. The
thing that's interesting to me, Adi, is we may not even know
when it has sentience, right? It's like we assume human-like
forms on intelligence. But if you read a lot of science fiction
like I do, maybe alien life is going to be carbon based, but
maybe not. Maybe they'll have a different metabolic
system, maybe different neural systems, and you need to be
ready for that. We may not even know it, that's the thing
ADI IGNATIUS:
This is fabulous. We're going to have to get you back on the
show because there's a lot more to talk about and we didn't
even get into the congressional hearings on aliens.
Karim Lakhani:
Oh gosh, yes.
ADI IGNATIUS:
This is just a half step away from that. So, this has been
Karim Lakhani, Harvard Business School professor. Karim,
thank you very much for being on the show.
Karim Lakhani:
Great to be here with you, Adi. Thank you.
READ MORE https://bit.ly/4moreeee