Great to see you all here, everyone having a good time?
I thought that that session was fantastic.
And that’s one of the things I love about Wired
is the interdisciplinary nature of the content.
So we’re moving from culture and entertainment,
now looking at business and technology.
And obviously as Katie has said this morning,
that one of the things that we are focused on at Wired
is figuring out what’s coming next,
and also how we build a better future.
And one of the things that we are really anticipating now
is the topic that’s dominating the conversation
we’re having about the future,
which is the deployment of AI.
Obviously we’re seeing leaps forward in the kind of modeling
and the release of consumer products
that have shifted the perception of what might be possible.
And that mass attention,
these rapid advances have had an impact on organizations,
both in the public and the private sector.
So I’m delighted to be joined this afternoon
by someone who has a really strong overview,
but also deep specific insights
into how organizations can integrate
and deploy artificial intelligence.
So please welcome Joe Atkinson,
Vice Chair and Chief Product and Technology Officer at PwC.
[upbeat music] [audience applauding]
Thanks. Nice to see you, Joe.
So Joe, great to see you. Thanks for being here.
Great to be here.
Super excited to have this conversation.
And so you’re talking to a lot of very large organizations.
You’re talking to a lot of boards
with a lot of responsibility on their shoulders.
They all recognize now,
obviously the technology is moving very, very quickly.
Things are moving very rapidly. They need to invest.
They don’t fall behind their competitors.
How do you think about they can best achieve that ROI
on that big capital deployment
they’re gonna have to make in order to compete?
So you’re gonna come right outta the gate
with the hard questions? Yeah, exactly.
What’s the ROI and how do we do that?
Yeah, no mercy.
So Greg, let me start with a huge thank you to you,
and Katie and the team.
‘Cause this has been just such a great privilege
to be with you all at the Wired conference,
and we’ve enjoyed every minute of it.
And I’m looking forward to the next minutes of it.
But now to your question.
This issue, in my 30 years with the firm,
which makes me sound old.
But in my 30 years with the firm,
I’ve never seen anything capture the C-suite
and the board agenda the way generative AI has.
There have been lots of topics of importance,
but nothing that has had this sweeping effect
where everybody wants to talk about it.
And the interesting thing is,
everybody is largely in the same boat.
Which is, they know it’s important.
In one of the conversations we were having,
we talked about the fear of missing out.
It’s probably the biggest FOMO
that you’ve ever seen in the C-suite ever.
Everybody wants to be ahead.
But there is a real concern about the cost
and the risk and the governance.
So the biggest issue that people struggle with
is where to start.
And I think the where to start question
is actually in some ways one of the simpler ones.
And that is start the same place you do any other investment
that you wanna make in your business or your organization.
Which is, Why am I investing?
And what’s the return I expect to get from it?
Now I’ll leave experimentation aside,
’cause we’ve had some really good conversations
in the hallways, in the sessions about experimentation,
and that’s really important.
But if you look at it on a pure business basis
and you say, Why would I invest?
It’s like any other investment.
You invest because you hope the value
that comes back from it
is a multiple of what you’ve put into it.
And in order to get that,
you have to have a perspective about what purpose
it’s going to serve in your organization,
what it’s going to deliver for you,
and what is the business outcome that you’re looking for.
And if you start with that framework,
and that’s what we’re encouraging boards to think about.
If you start with that framework,
it can take you in a pretty logical direction from there.
Sure. Things are moving so quickly though.
It’s very, very hard for boards to actually keep pace
with what’s happening to actually understand
but also that we have to kind of think about
how the challenges of this come about.
Do you see more and more a trend
towards boards appointing a Chief AI Officer,
someone who actually has deep knowledge, deep experience,
and can offer advice to board level, to the C-suite?
Yeah, we’re definitely seeing an interest
in having somebody in the C-suite that has accountability
and responsibility for the AI strategy.
Whether it’s a Chief AI Officer, or whether it’s the CTO.
And it depends a little bit on how the organization
is already organized.
Many organizations worry about adding another seat
to the C-suite, but yet,
because this conversation has been so profoundly impactful
in terms of business models and where it’s going,
it is definitely something we’re seeing.
I think the other point that’s in your question
that is also if not as important, maybe even more important,
which is this question about the speed
with which this is all traveling,
and how do you get the confidence that you need,
that you’re evaluating the opportunity
and you’re evaluating the risks the right way.
That conversation is happening in every organization,
and they’re looking for help across their organization
on how to do that.
We were talking earlier in the hallway
about just the speed of up-skilling
and the challenge that people have
in keeping pace on technology.
And most of us that have spent our careers
in any degree of technology would say
it’s always been hard to keep pace.
There’s always some innovation happening
that is interesting, that’s new.
I live in the world where my CEO likes to send me emails
or texts at four in the morning and ask questions like,
Have you seen this startup?
Did you see this investment? Or where are we on this?
I think we all live that. And I have bad news.
It’s just gotten remarkably worse in the generative AI world
and in the AI world.
The speed just keeps feeling like it’s picking up.
So one of the things we’ve been sharing with clients is,
and particularly C-suites and boards is, Get comfortable.
This phrase is probably overused, but I think is important,
get comfortable being uncomfortable.
It isn’t going to get any better.
You’re going to have to get comfortable with the fact
that you are behind the eight ball every day
when you wake up in the morning,
and the innovations are going faster
than any of us can reasonably keep up on.
So what’s the solution to that?
The solution to that is make sure you’re surrounded
by really smart people.
Make sure you’re giving people that opportunity
to experiment and innovate.
Make sure you’re putting those investments on the plate
so that you have an opportunity
to learn from your own organization
and you have a mechanism to bring it in from others.
And that idea of being comfortable
with being uncomfortable, I think,
is something that we’ve heard a lot in technology
in the last few years in the sense of,
you have to be able to fail.
Like when you are working with a board,
they have a fiduciary duty to deliver
for their shareholders.
How do you get that balance right
between being willing to fail, to experiment,
to try things,
but also having to deliver on this duty that they have?
Yeah. Legally binding duty.
The good news is I can’t see most of the faces
in the audience.
And so before I say this,
I’ll hope that none of our board member clients
are in the room when I say what I’m about to say.
But none of them understand anything.
Like, they’re all facing this massive challenge
in the speed of innovation.
And the good news is, the really good news is,
the really good boards know they don’t understand.
So the first question has gotta be, Help me understand.
If that conversation isn’t taking place in the C-suite,
if it’s not taking place in the boardroom,
then I think the challenges get worse.
But if the acknowledgement is that
in a world where we’re all uncomfortable all the time,
and the pace is moving so quickly,
if you’re willing to acknowledge that,
you have to be in the sense,
equally willing to acknowledge that the learning
is gonna come from somebody else.
It’s gonna come from the innovators in the space,
it’s gonna come from startup communities,
it’s gonna come from academia.
It’s gonna come from places
that maybe we’re not as good collectively
as executives and leaders
of constantly monitoring and engaging.
But that’s what it’s going to require in this world.
So thinking about sort of like how organizations adapt,
they’re gonna have to start thinking about regulation.
It’s coming down the pipe.
We know it’s something that’s real.
We’ve seen voluntary commitments
from a number of organizations, going to the White House,
and making these kind of voluntary promises.
What’s your view on the current regulatory environment?
And how is that gonna shift moving forward?
So Greg promised no hard questions
when we prepared for this session,
and now we’re asking me to give you a sense
on the regulatory environment.
What could possibly go wrong?
I would love to be able to predict
the regulatory environment.
And there are a couple things I think, actually,
that are showing themselves as likely.
First you mentioned the voluntary commitments
over the summer.
Several of the largest technology companies
with large language models and innovations and AI,
came together at the White House’s urging.
Came out with some voluntary standards,
really important setting.
That became the foundation
for the President’s executive order
that was released at the end of October.
If you haven’t read the executive order,
I really encourage you to print it out.
‘Cause that part’s fun.
You get 100-plus pages on the nightstand.
And then if you need something to help you sleep,
you can read the entirety of the order.
But there is a lot of important learnings in the order.
And one of them, I think,
is this issue that a lot of people look
to the voluntary standards and said,
Okay, this is about large technology companies.
The order extends that question.
It extends it to people
that are in critical infrastructure businesses.
Think healthcare and transportation, airlines, et cetera.
It also has a lot of discussion on cybersecurity.
Yeah. And we all,
I think, in this room, understand the implications
of generative AI on cybersecurity,
but clearly the White House does as well,
as they spend a lot of the real estate
of the order talking about cybersecurity.
So I think it’s really important, again,
probably the same theme to get educated,
make sure somebody’s monitoring this.
Predictively, I don’t think we’re gonna see
any meaningful regulatory change coming outta Washington
until we get our way through the presidential election.
That actually is probably, may sound like good news,
but it’s probably bad news.
Because what that will likely mean
is you will see state regulation developing.
In fact you’re already seeing it in California,
New York and elsewhere. Yeah.
And that state regulation means
for those of us that operate generally,
across multiple states,
you’re gonna have to digest lots
of different requirements and obligations.
And of course we’re just talking about the States right now.
Go to the Bletchley Park meetings
that were held not long after
the President’s executive order,
and there was a lot of discussion about
what the rules should be in the EU and globally.
So I think we’re gonna be dealing
with a very, very complicated regulatory environment.
But I do think there are some really good lessons
to take away from the executive order
because it lays out a lot of the principles
that I think we’re gonna be talking about
for the regulatory environment for the coming years.
Yeah. I mean it’s interesting that clearly Congress
has yet to even pass any meaningful legislation
around the platforms of Web 2.0.
So the idea that we’re kind of gonna be able to keep up
with what’s happening in generative AI,
when as a publication, Wired,
just keeping up with the day-to-day news
on what’s happening is pretty tough.
Like, I’m interested to kind of get your sense
of that speed of innovation and how lawmakers,
and I guess it’s the same thing you were talking about
with people in the C-suite.
Like, how to sort of try and stay ahead of that?
It’s a very hard sort of like,
it’s almost like we need technologists and government,
we need technologists at the highest level of organizations
to be able to kind of act as translators, almost.
Yeah, I think that’s such a great point.
And I’d come back to something that I say a lot.
Which is, there used to be jobs that were technology jobs,
and then there were non-technology jobs.
It’s just not a meaningful difference anymore,
And I hadn’t often included our regulators
and our policy makers in that,
but there used to be non-technology policy makers
and technology policy makers.
And now it’s just policy makers.
So I think the challenge for anyone, is again,
that theme of recognizing the challenge that we all face,
which is climbing a very steep
and rapidly accelerating learning curve.
But the second piece that is maybe good news,
’cause I’m an optimist at heart.
The good news is the regulatory environment
has been engaging in this discussion.
The regulatory, the lawmakers, policymakers,
they’re engaging in this discussion.
And they’re engaging to learn.
There’s exceptions to every rule, by the way.
But most of them are engaging to learn.
And I think that’s really important.
So part of the recommendation,
and it goes to not just the policymakers,
but it goes to organizationally is,
the kinds of industry organizations
that most of us belong to, the professional organizations,
sometimes they’re lobbying organizations,
but sometimes they’re just simply trade groups.
Those trade groups are gonna play a really important role
in engaging with policy makers.
And many of us might have observed those
from a distance in the past.
I think it’s important for us to engage really closely
with the trade groups and the associations
that many of us are affiliated with,
because they will help shape that knowledge
that’s gonna be so important to the policy makers.
And we’ll of course the regulators and policy makers
aren’t gonna get everything right.
But I think the reality is, we’re gonna look at a lot of,
kind of Matrix-level complexity
for the coming months and years.
Yeah, I think maybe we’re gonna learn some of the lessons
that we learned from the platforms of Web 2.0,
and actually understand that we are all participants in this
and we can all help to shape what the future looks like
around this kind of technology.
Obviously one of the conversations
that we’re having around regulation
is about the future of work.
The concern that we’re gonna see a lot of people finding it,
a lot of the tasks they perform are gonna become automated.
I’m interested to get your sense on how you feel
that the kind of the tools that we’re now seeing?
These consumer products that are out there
in the marketplace
that have really changed the conversation.
How you think they’re gonna affect the way
that work gets done in the coming years?
You’re always tempted when somebody gives you a microphone
to speak in massive hyperboles,
and then you try to make sure you don’t go too far.
So you all can judge me on this,
but I think the change is profound.
And I think we’re gonna be living with it
for decades to come,
in terms of the nature of work and what will change.
And I’ll give you an example from our own business.
Maybe not a business many of you think much about,
but the professional services businesses at large.
Accounting, consulting, legal,
they will change dramatically.
So some of you may have started your career
in those businesses.
And when you start your career in those businesses,
generally speaking, your first year or two,
or three or five,
depending on what part of the business you join,
they’re a bit of a grind.
We have a model where we really immerse you in the work
in order to help you learn the arena.
And a lot of the work that today, well not even today,
a lot of the work that I did is massively different.
But the work that happens today
is going to be even more massively different
because much of it’s going to go away.
And so the challenge I think,
for professional services firms like ours,
and for organizations in general will be,
how do I equip people with knowledge?
Because if you believe as I do
that the large language models are largely commodities
or will become commodities over time,
then the knowledge of individuals and organizations
becomes where value can be derived.
Well, in order for organizations to have individuals
with a lot of knowledge, they’ve gotta learn things.
They’ve gotta build that up.
And if you take away some of the activity
that helped create that learning
in early stage career development, then collectively,
I think we have to figure out how to replace that.
What does that look like?
And it might just be a different way of learning.
And fundamentally though,
this is not a question about work,
it’s a question about tasks, right?
Because work’s not going away.
Yeah. But tasks will change.
Do you think that’s accurate?
I do think that’s accurate.
And people talk a lot about job losses
as this innovation takes place.
And we all should be aware and monitoring that.
But the reality is, it’s tasks.
Yeah. Tasks get disrupted.
And each of our jobs is made up of some collection of tasks.
If your job happens to be made up of all of tasks
that are gonna get disrupted,
then the question then becomes,
What is the new promise, opportunity for that individual,
and for people that are doing that job?
And this is where I think the up-skilling question,
and again, we’ve all been talking up-skilling
for many years,
but I think it’s probably even a more critical moment now
than what we’ve seen in the last 10 years.
Which is, if you believe as I do that organizations
that employ people have some obligations to those people,
that you can’t just hire somebody, wish them luck,
and if things don’t work out, send them off on the street.
If you believe that there’s an obligation broader than that,
then I think organizations have to help people
cross the bridge.
If today their role and their tasks
are going to be disrupted by technology, innovation,
whether it’s generative AI or anything else for that matter,
then the question is,
Do I have a talented employee?
Or do I not have a talented employee?
If I have a talented employee,
then I wanna help them cross the bridge.
I wanna help them get to whatever the next capability
and contribution is.
But there are also places where I think
this has fundamental opportunities for businesses
and I’ll come back again to ours ’cause I live it every day.
If I ask for a show of hands,
how many of your kids are going into accounting?
You don’t have to raise your hands.
I suspect most of you are not telling your kids
to go into accounting.
I will tell you it’s not a bad profession,
if your kids are thinking about it.
But the reality is that probably about 15 years ago,
the profession for a lot of good reasons,
decided that you needed a Master’s degree
to get a Certified Public Accounting certification.
And what has happened is, as kids are looking at college,
and they’re saying,
Okay, first of all, I’m getting saddled with debt.
Do I really need a Master’s degree?
Do I wanna go to school for,
it ends up usually being a five-year combined undergrad
And I wanna spend that extra money
if only to enter a profession that’s gonna work me
really, really hard for a few years.
Do I need to do that? Or do I just go direct?
That’s created a constraint,
not only for services firms like ours,
but for our clients and the finance functions
and the accounting functions.
There are, I know this is shocking,
but there will not be enough accountants.
And so for us,
generative AI creates an opportunity to open up
this profession to not only more people,
but also to address the practical reality
that our clients are gonna need that advice,
It might take a different form in the future
than it did in the past,
but they’re gonna need that perspective.
And I may not have enough bodies, if you will,
coming into the profession.
This gives me a way to solve that in a different way.
All right, you heard it here first.
The great accountancy crisis of 2030 is coming guys,
So I think that, on that point actually,
we hear a lot about sort of like AI impacting,
sort of like low-skilled tasks.
So taxi drivers, retail.
I saw a really interesting thing recently
was that Goldman Sachs used to have 500 people
on their equities trading desk.
They now have three.
So like, I’m interested in digging into
what you were saying about up-skilling.
I think Goldman Sachs is gonna be absolutely fine, right?
Is it just the case that we’re shifting the kind,
we’re creating new jobs,
that there is this kind of very folksy belief, right?
That there’s only so much work out there,
but actually we’ve seen through multiple sort
of like shifts in technology
and multiple shifts in our economies that actually,
you can grow the number of jobs out there.
It is possible to employ more people.
How do you think about that kind of, that big picture,
in the sense of like work isn’t gonna go away, right?
We’re just gonna be doing different things.
Yeah, again, I’ll come back to the optimistic point.
The optimistic point is work isn’t gonna go away.
There’ll be more jobs than less.
And all the studies by the way,
tell us that anybody that’s looked at it mathematically
comes to that conclusion.
But if you come back to another point that you hit
that I think is so important, Greg, which is,
if you look historically at technology innovation,
historically, every major technology innovation
created a period of growth,
or opened up opportunities for growth.
You really can’t go back in history
and find a technology innovation
that lowered the human condition.
Now, there’s a transition moment,
there’s transition period
and some people are impacted more heavily
in the direct way than others.
So I don’t mean to be callous about that.
Clearly there will be impact.
But the opportunity actually will create openings, I think,
for different ways to do things
and different ways to think about things.
And I’ll zero in on two examples.
And it goes to your point about whether
this will all be the low-value tasks.
I think it’s coming first for the low-value tasks
and I think most of us would be happy about that
because I do plenty of low-value tasks in the day
that I’d be happy to have a generative AI tool do instead.
But it’s also coming for knowledge work in general.
It’s coming for lawyers, accountants,
professional service providers and consultants.
It’s coming for finance professionals,
it’s coming for analysts,
it’s coming for all of us in one way or another.
And that is okay.
It is the opportunity that it creates
to think about work differently.
So when we’re all doing automation
and a lot of the automation and data opportunities
were showing themselves in the pre-pandemic years,
everybody worried about what does this all mean
that we can get to the four-day work week.
And I don’t know that we worried about that so much
as we thought that that was a really good idea.
I do think that there is an opportunity
to reshape the fundamentals of the way
that we all work in ways that were simply not available
to us 10 years ago, 20 years ago, 30 years ago.
I think that in some ways is the burden of this generation
of leadership in C-suites and venture capital,
in talent organizations, et cetera, is,
are we willing to challenge the constructs sufficiently
to think about work in a different way?
Are we gonna learn the lessons of the pandemic
and hybrid work and all the other things
and experiences that we’ve had,
and apply it in meaningful ways to envision a new way,
a truly new way of working?
So we’ve talked a lot about large organizations.
I’m keen to get your sense also though,
like we’ve got a lot of entrepreneurs,
a lot of startups in the audience.
Do you see these new tools are gonna lower the barriers
to access, for new gen of entrepreneurs,
that they can get way further down the road
than they were able to previously,
by using these kinds of tools?
There’s no doubt in my mind that it will open up
a new period of innovation and entrepreneurialism.
And anybody that has lived in the startup community,
and I’ve spent 30 years with the firm,
but I worked with some startups in college.
So I lived that world early on.
Now that was different startups, ’cause it was 30 years ago.
They weren’t doing anything nearly as cool
as probably some of the things happening in this room.
But that opportunity meant that at the time,
and this is still largely true,
you need somebody thinking about the financials,
you need somebody thinking about the coding.
You need somebody thinking about how to test the platform.
And so much of that will now be opened up
with access to generative AI and other tools
that you get this phenomenon of the one-person startup.
We talk a lot about the one-person startup
and this opportunity that these tools open up for people
to innovate with a lot less investment on the front end.
Which if you’re in the venture capital space,
or you’re in the space where you’re thinking about
angel investing or what have you,
it also means that you can get further in the premise
before you have to commit the kind of capital
that’s required for scale.
And that’ll take a little while to work out.
And that doesn’t happen overnight,
but I just think it creates a lot of opportunity.
Yeah, and I think that one of the interesting things
is we see in like a lot of large Western economies,
like there’s been a drop off in productivity.
Like is generative AI gonna, or any form of AI,
large language models, whatever,
are you gonna see machine learning
help kind of boost that productivity
because there’s only so many levers we can pull
in order to grow an economy.
Yeah. It’s funny.
I was having this conversation with somebody over lunch.
The idea of productivity
and what we all lived in the pandemic,
which was like productivity was my waking hours
were available to anybody that pinged me on my laptop
or my phone.
Hopefully that model goes away.
If we trade a little bit of productivity for that,
I think I would sleep okay at night.
But I do think that if you go to the most simplistic view,
which I love a good simplistic view.
You say, Today, I spent some portion of my time,
five minutes, two hours, four hours.
I spent some portion of my time doing tasks
that are gonna get replaced.
Then the question is,
What do I do with the energy that that unlocks in me?
In my own day?
And I can dedicate it to family,
I can dedicate it to friends, I can pick up hobbies,
I can spend more time at the office
or spend more time on the work.
But wouldn’t it be great to be empowered
to make that decision?
And I think that again, is the promise here.
Okay, so we’re nearly at the end of our time.
Yeah. But I need some advice.
So imagine you’re a VC. Okay.
And you are investing for your LLPs.
Where are you gonna put your money in this next period?
Where are we gonna see the real growth?
Where are we gonna see the real opportunities?
Yeah, the funny part is,
which will sound like a non-answer.
I think the opportunities are so significant
in so many places. Right.
That I think it’s gonna come down to the same way
that we’ve made investment decisions for years.
Which is, Where do I wanna play?
So if I’m the venture capitalist,
I have a couple billion dollars in my pocket.
So where would I would play?
I realize not all the venture capitalists
have a couple billion, by the way.
But for me, healthcare.
I’ve talked to a number of people over the last few hours
Healthcare has massive opportunities
to rethink those businesses, the nonprofit communities,
Financial services is gonna be massively impacted by this.
All of the consumer grade technology that is showing itself
is so powerful in the financial services industries today.
It’s the beginning stages of that, in many ways.
And you look at the retail and the consumer relationships
and what that looks like.
A few years ago we’re all saying it was the death of retail,
but now you’re seeing a very different
retail value proposition.
So I think new retail is kind of an interesting one.
So in some ways, again, in the technology, the media space,
there’s a ton of ’em.
So it’s a good time to be a venture capitalist.
So basically, we can’t go wrong.
I don’t think you can, honestly.
I think apply the smart decisions as usual.
A little bit of discipline goes a long way
in the venture capital space.
But I don’t think you can go wrong.
All right. Terms and conditions apply.
But Joe, I’m gonna have a word with Steven Levy afterwards,
’cause I think you should be made an honorary member
of Steven’s AI Optimism Club after that talk.
Thank you so much for being with us.
Thank you. Appreciate it, Greg. Thank you so much.
Thanks mate. Enjoyed that.