Demis Hassabis On The Future of Work in the Age of AI

[00:00] - It's a very intense time in the field.

[00:02] We obviously want all

[00:03] of the brilliant things

[00:05] come up with new cures for

[00:07] incredible things for humanity.

[00:09] That's the promise of AI.

[00:11] But also, there are worries

[00:12] if the first AI systems are built

[00:13] with the wrong value systems

[00:16] that could be also very bad.

[00:18] - Wired sat down with Demis Hassabis,

[00:20] who's the CEO of Google

[00:23] of the company's artificial intelligence.

[00:25] He's a Nobel Prize

[00:27] We discussed AGI, the future of work,

[00:30] and how Google plans to

[00:33] This is "The Big Interview."

[00:35] [upbeat music]

[00:42] Well, welcome to "The

[00:43] - Thank you, thanks for having me.

[00:44] - So let's start talking

[00:48] Now, you founded DeepMind with the idea

[00:51] that you would solve intelligence

[00:56] to solve everything else.

[00:57] And I think it was like a 20-year mission.

[00:59] We're like 15 years into

[01:02] - I feel like, yeah,

[01:02] we're pretty much dead on track, actually,

[01:04] is what would be our estimate.

[01:06] - That means five years away

[01:08] from what I guess people will call AGI.

[01:11] - Yeah, I think in the

[01:13] that would be maybe 50% chance

[01:16] that we'll have what we

[01:18] - Well, some of your peers are saying,

[01:20] "Two years, three years,"

[01:22] and others say a little more,

[01:25] that's really soon.

[01:27] How do we know that we're that close?

[01:30] - There's a bit of a debate

[01:32] in the field about definitions of AGI,

[01:34] and then obviously, of

[01:36] There's different predictions

[01:39] We've been pretty consistent

[01:41] And actually, Shane Legg,

[01:42] one of my co-founders

[01:44] you know, he helped define

[01:47] early 2001 type of timeframe.

[01:50] And we've always thought

[01:52] that has the ability to exhibit,

[01:54] sort of all the cognitive

[01:58] And the reason that's important,

[01:59] the reference to the human mind,

[02:01] is the human mind is the

[02:04] Maybe in the universe, the

[02:07] So if you want to claim sort

[02:10] then you need to show that it generalizes

[02:13] to all these domains.

[02:14] - Is when everything's filled in,

[02:15] all the check marks are

[02:20] - Yes, so I think there are

[02:23] You know, that all of us

[02:24] who have used the latest

[02:27] will know very well, like on reasoning,

[02:29] on planning, on memory.

[02:30] I don't think today's

[02:33] do true invention,

[02:35] you know, true creativity,

[02:36] hypothesize new scientific theories.

[02:39] They're extremely useful,

[02:41] but they have holes.

[02:43] And actually, one of the

[02:45] we are at AGI yet is

[02:47] because of the consistency of responses.

[02:50] You know, in some domains,

[02:52] we have systems that can do

[02:55] math problems to gold medal standard-

[02:57] - Sure.

[02:59] But on the other hand,

[03:00] these systems sometimes still

[03:03] or even counting the number

[03:05] - Yeah.

[03:07] what you would expect.

[03:08] That level of sort of difference

[03:10] in performance across the

[03:14] and therefore shows

[03:15] that these systems are not

[03:17] - But when we get it,

[03:18] is it then like a phase

[03:21] then all of a sudden things are different,

[03:24] all the check marks are checked?

[03:26] - Yeah.

[03:27] and we have a thing

[03:28] - Mm-hmm.

[03:29] - Are we then power in a new world?

[03:30] - I think, you know, that again,

[03:32] that is debated, and it's not clear to me

[03:34] whether it's gonna be more

[03:35] of a kind of incremental

[03:41] My guess is, it looks

[03:43] of an incremental shift.

[03:44] Even if you had a system like

[03:47] still operates with the physical laws,

[03:50] you know, factories,

[03:53] So it'll take a while for the

[03:56] this sort of digital

[03:58] to really impact, I think, a

[04:02] Maybe another decade plus,

[04:04] but there's other theories on that too,

[04:05] where it could come faster.

[04:06] - Yeah, Eric Schmidt, who I

[04:10] has said that, "It's almost

[04:12] He says, "If China,

[04:16] then we're cooked."

[04:17] Because if someone gets

[04:19] before the next guy, then

[04:24] You know, because then

[04:26] bigger leads there.

[04:27] You don't buy that, I guess.

[04:29] - I think it's an unknown.

[04:30] It's one of the many unknowns,

[04:31] which is that, you know,

[04:32] that's sometimes called

[04:34] where the idea there is

[04:37] they're able to self-improve,

[04:39] maybe code themselves

[04:41] that maybe they're extremely

[04:43] So what would be a slight lead,

[04:46] let's say, you know, a few days,

[04:48] could suddenly become a

[04:51] But there are many other

[04:53] where it's more incremental.

[04:54] Some of these self-improvement

[04:57] to kind of accelerate in that way,

[05:00] then being around the same time,

[05:03] would not make much difference.

[05:05] But it's important, I mean,

[05:06] these issues are the geopolitical issues.

[05:08] I think the systems that are being built,

[05:10] they'll have some imprint of the values

[05:13] and the kind of norms of the

[05:16] that they were embedded in.

[05:18] - So, you know, I think it is important,

[05:20] these kinds of international questions.

[05:22] - So when you build AI at Google,

[05:26] you know, you have that in mind.

[05:28] Do you feel competitive imperative

[05:32] "Oh my God, we better be first?"

[05:34] - It's a very intense time

[05:37] as everyone knows.

[05:38] There's so many resources going

[05:41] lots of things that need to be researched.

[05:43] And there's sort of

[05:45] of pressures going on.

[05:46] We obviously want all

[05:48] that these AI systems can do.

[05:50] You know, I think eventually,

[05:51] we'll be able to advance

[05:54] like we've done with AlphaFold,

[05:56] come up with new cures for

[05:58] incredible things for humanity,

[06:02] But also there are worries

[06:05] if the first AI systems are built

[06:07] with the wrong value systems

[06:10] that could be also very bad.

[06:11] And, you know, there

[06:14] that I worry a lot about.

[06:15] One is, bad actors in

[06:17] or rogue nations repurposing

[06:21] for harmful lens.

[06:22] And then the second one is, obviously,

[06:24] the technical risk of AI itself.

[06:26] As it gets more and more powerful,

[06:27] more and more agentic,

[06:28] can we make sure the

[06:32] They can't be circumvented.

[06:33] And that interacts with

[06:36] what are the first systems that are built

[06:38] by humanity gonna be like?

[06:40] There's commercial imperative-

[06:42] - [Steven] Right.

[06:43] and there's a safety aspect to worry

[06:46] about who's in the lead and

[06:50] - A few years ago, the

[06:53] "Please, regulate us.

[06:54] We need regulation."

[06:55] - And now, in the US at least,

[06:57] the current administration

[07:00] in putting regulations on

[07:04] so we can beat the Chinese.

[07:07] Are you still asking for regulation?

[07:09] Do you think that that's

[07:11] - I think, you know,

[07:13] and I've been consistent in this,

[07:14] I think there are these other

[07:19] that have to be taken into account,

[07:21] and the world's a very different place

[07:22] to how it was five years

[07:25] But there's also, you know,

[07:26] I think the idea of smart regulation

[07:29] that makes sense around these

[07:32] I think is gonna be important.

[07:33] I continue to believe that.

[07:35] I think though, and I've

[07:37] it sort of needs to be international,

[07:38] which looks hard at the moment

[07:40] in the way the world is working,

[07:41] because these systems, you know,

[07:43] they're gonna affect everyone,

[07:45] and they're digital systems.

[07:47] - So, you know, if you sort

[07:51] that doesn't really help

[07:52] in terms of the overall safety

[07:53] of these systems getting

[07:57] and as a society.

[07:58] - So that's the bigger problem, I think,

[08:00] is some kind of international

[08:03] I think, is what's required.

[08:05] And then smart regulation,

[08:07] that moves as the knowledge

[08:09] about the research

[08:13] - Would it ever reach a point

[08:15] "Man, we're not putting the guardrails in.

[08:18] You know, we're competing,

[08:21] or you can't get involved in that?"

[08:24] - I think a lot of the

[08:27] at least the western labs,

[08:29] you know, there's a small number of them

[08:31] and we do all know each other

[08:33] and talk to each other regularly.

[08:34] And a lot of the lead researchers do.

[08:35] The problem is, is that it's not clear

[08:39] we have the right definitions

[08:42] Like, today's systems,

[08:43] although they're impressive

[08:46] they're also very flawed.

[08:47] And I don't think today's systems,

[08:49] are posing any sort of existential risk.

[08:52] - Mm-hmm.

[08:55] but the problem is that a lot of unknowns,

[08:56] we don't know how fast those will come,

[08:58] and we don't know how risky they will be.

[09:00] But in my view, when there

[09:03] then I'm optimistic we'll overcome them.

[09:06] At least technically,

[09:07] I think the geopolitical

[09:09] end up being trickier, given

[09:12] and thoughtfulness, you know,

[09:13] sort of using the scientific method

[09:15] as we approach this AGI point.

[09:18] - That makes perfect sense.

[09:20] But on the other hand, if

[09:23] we just don't have much time, you know?

[09:25] - No, we don't.

[09:26] We don't have much time.

[09:27] I mean, we're increasingly

[09:31] and things like cyber,

[09:34] and also research into controllability

[09:37] and understanding of these systems,

[09:38] sometimes called mechanistic

[09:40] You know, there's a lot of

[09:43] - Yeah, that's right.

[09:44] I wanna get to interpretability.

[09:45] - Yeah, that are being invested in,

[09:46] and I think even more needs to happen.

[09:48] And then at the same time,

[09:50] we need to also have societal debates more

[09:53] about institutional building.

[09:55] How do we want governance to work?

[09:57] How are we gonna get

[09:59] at least on some basic principles,

[10:01] around how these systems

[10:04] and also built?

[10:05] - What about the effect on

[10:09] - Yeah.

[10:10] how much do you feel that AI is going

[10:13] to change people's jobs,

[10:15] you know, the way jobs are

[10:18] - I don't think we've seen,

[10:19] my view is if you talk to economists,

[10:20] they feel like there's

[10:23] You know, people are

[10:25] certainly in certain domains-

[10:26] - [Steven] Yeah.

[10:27] many, many scientists are using

[10:30] So it seems to be additive at the moment.

[10:32] We'll see what happens over

[10:34] I think there's gonna be a lot of change

[10:37] with the jobs world, but

[10:40] what generally tends to

[10:44] that are actually better,

[10:45] that utilize these tools

[10:47] what happened with the internet,

[10:49] We'll see if it's different this time.

[10:51] - Yeah.

[10:52] - Obviously everyone

[10:53] will be different.

[10:53] And it may be, it will be,

[10:55] but I think for the next few years,

[10:57] it's most likely to be, you know,

[10:59] we'll have these incredible tools

[11:00] that supercharge our productivity,

[11:03] make us really useful for creative tools,

[11:07] and actually almost make

[11:09] in some ways in what we're

[11:13] So I think there's gonna

[11:16] over the next period of

[11:19] - Well, if AGI can do

[11:22] then it would seem that they

[11:24] - That's the next question

[11:28] But, you know, even if you

[11:30] there's a lot of things I

[11:32] with a machine.

[11:34] You know, I sometimes give this example

[11:36] of doctors and nurses.

[11:38] You know, maybe a doctor

[11:39] and what the doctor

[11:41] you know, one could imagine

[11:44] or even having an AI kind of doctor.

[11:47] On the other hand, like nursing,

[11:49] you know, I don't think you'd

[11:51] I think there's something

[11:52] about the human empathy

[11:55] and so on, that's particularly humanistic.

[11:59] I think there's lots of examples like that

[12:01] but it's gonna be a

[12:04] - If you would talk to a graduate now,

[12:07] what advice would you give

[12:09] to keep working-

[12:11] - Through the course

[12:12] of a lifetime-

[12:14] - You know, in the age of AGI?

[12:16] - My view is, currently,

[12:18] and of course, this is

[12:20] with the technology developing.

[12:22] But right now, you know,

[12:24] if you think of the next

[12:27] the most productive people

[12:30] if they are native with these tools.

[12:33] So I think kids today, students today,

[12:36] my encouragement would be immerse yourself

[12:39] in these new systems, understand them.

[12:41] So I think it's still important

[12:43] to study STEM and

[12:45] so that you understand how they're built,

[12:47] maybe you can modify them yourself

[12:49] on top of the models that are available.

[12:50] There's lots of great open

[12:53] And then become, you know,

[12:55] incredible at things like

[12:59] you know, system instructions,

[13:01] all of these additional

[13:03] And really know how to get

[13:07] and do it for your

[13:10] and things that you are

[13:12] And then come out of that being incredible

[13:14] at utilizing those new tools

[13:17] for whatever it is you're going to do.

[13:18] - Let's look a little beyond

[13:21] Tell me what you envision

[13:26] in 20 years, in 30 years,

[13:30] what's the world like

[13:33] - Well, if everything goes well,

[13:34] then we should be in an era of what I like

[13:37] to call sort of radical abundance.

[13:40] So, you know, AGI solves

[13:43] what I sometimes call root node problems

[13:45] in the world facing society.

[13:46] So a good one, examples

[13:50] much healthier, longer lifespans,

[13:52] finding new energy sources,

[13:54] you know, whether that's optimal batteries

[13:56] and better room temperature,

[14:01] And then if that all happens,

[14:04] then we know it should be a kind of era

[14:06] of maximum human flourishing

[14:10] and colonize the galaxy.

[14:14] You know, I think the

[14:16] in the next 20, 30 years if

[14:20] - I'm a little skeptical of that.

[14:22] I think we have an

[14:25] but we don't distribute it,

[14:27] you know, fairly.

[14:28] - I think that we kind of know

[14:29] how to fix climate change, right?

[14:31] We don't need a AGI to

[14:33] yet we're not doing it.

[14:36] I think we being as a species,

[14:38] a society not good at collaborating,

[14:41] and I think climate is a good example.

[14:42] But I think we are still operating,

[14:45] humans are still operating

[14:48] Because actually, the

[14:50] relative to the amount

[14:53] in our cities.

[14:54] And I mean, this is why

[14:56] are being destroyed,

[14:58] and it's affecting

[15:01] and everything.

[15:02] - And it's also partly

[15:04] to accept, we do now

[15:08] But it would require

[15:10] - Yeah.

[15:12] But this radical abundance

[15:15] We would be in a finally, like,

[15:17] it would feel like a non-zero-sum game.

[15:20] - How will we get [indistinct] to that?

[15:21] Like, you talk about diseases-

[15:22] - Well, I gave you an example.

[15:24] and now some people think

[15:26] - Let me give you a very simple example.

[15:28] - Sure.

[15:29] This is gonna be a huge issue

[15:31] It's already an issue.

[15:32] Countries in different, you know,

[15:34] poorer parts of the world,

[15:35] also obviously compounded

[15:38] - [Steven] Yeah.

[15:38] - We have a solution to water access.

[15:40] It's desalination, it's easy.

[15:42] There's plenty of sea water.

[15:43] - Almost all countries have a coastline.

[15:45] But the problem is, it's salty water,

[15:47] but desalination only very rich countries.

[15:50] Some countries do do

[15:52] as a solution to their

[15:54] but it costs a lot of energy.

[15:55] - But if energy was essentially zero,

[15:57] there was renewable free

[16:01] Like fusion, suddenly, you

[16:04] Water is, who controls a river

[16:06] or what you do with that does not,

[16:08] it becomes much less

[16:11] I think things like water access,

[16:13] you know, if you run forward 20 years,

[16:15] and there isn't a solution

[16:17] to all sorts of conflicts,

[16:18] probably that's the way it's trending-

[16:19] - Mm-hmm, right.

[16:21] further climate change.

[16:22] - So-

[16:23] many examples like that.

[16:23] You could create rocket fuel easily-

[16:25] - Mm-hmm.

[16:27] that from seawater, hydrogen and oxygen.

[16:29] It's just energy again.

[16:30] - So you feel that these problems

[16:37] then we're going to,

[16:41] and we will be-

[16:44] Yes, that's what I hope.

[16:45] But that's still a secondary part.

[16:47] So the AGI will give us the

[16:50] technically, like the water access.

[16:52] - Yeah.

[16:53] and this is where I think we

[16:55] or social scientists to be involved.

[16:58] That should hopefully shift our mindset

[17:01] as a society to non-zero-sum.

[17:04] You know, there's still the issue

[17:05] of do you divide even the

[17:08] Of course, that's what should happen.

[17:10] But I think there's much more likely,

[17:11] once people start

[17:13] that there is this almost

[17:18] and energy and things like that.

[17:19] - Do you think that

[17:23] by profit-making companies

[17:26] We're most likely to reach

[17:27] that optimistic high point through that?

[17:29] - I think it's the current capitalism

[17:31] or, you know, is the current

[17:32] or the western sort of

[17:37] have so far been proven

[17:39] to be sort of the best

[17:41] - Mm-hmm.

[17:43] My view is that once you get

[17:44] to that sort of stage of

[17:48] I think economics starts changing,

[17:51] even the notion of value and money.

[17:53] And so again, I think we need,

[17:55] I'm not sure why economists

[17:57] if maybe they don't believe

[18:00] But if they really did that,

[18:04] then I think there's a lot

[18:06] of economic new economic

[18:08] - You know, one final thing,

[18:10] I actually agree with you

[18:14] and is gonna have a huge impact.

[18:16] But when I write about it,

[18:18] I always get a lot of response from people

[18:21] who are really angry already

[18:26] and what's happening.

[18:28] Have you tasted that?

[18:29] Have you gotten that pushback

[18:34] It's almost like the

[18:36] - Yeah.

[18:37] - I mean, I think that anytime there's,

[18:39] I haven't personally seen a lot of that,

[18:41] but obviously, I've read

[18:42] and it's very understandable.

[18:44] That's all that's happened many times.

[18:45] As you say, industrial revolution,

[18:47] when there's big change,

[18:48] a big revolution.

[18:49] - And I think this will be at least

[18:50] as big as the industrial

[18:52] That's surprising, there's unknowns,

[18:54] it's scary, things will change.

[18:56] But on the other hand,

[18:57] when I talk to people about the passion,

[18:59] the why I'm building AI-

[19:00] - Which is to advance science

[19:01] and medicine-

[19:02] - And understanding of

[19:04] And then I explain to people, you know,

[19:06] and I've demonstrated, it's not just talk.

[19:08] Here's AlphaFold, you know,

[19:09] Nobel Prize winning breakthrough,

[19:11] can help with medicine and drug discovery.

[19:13] Obviously, we're doing

[19:14] to extend it into drug discovery,

[19:16] and we can cure terrible diseases

[19:18] that might be afflicting your family.

[19:20] Suddenly, people are like,

[19:21] "Well, of course, we need that."

[19:23] - Right.

[19:24] to have that if that's within our grasp.

[19:26] And the same with climate and energy.

[19:29] - Yeah.

[19:30] many of the big societal problems,

[19:31] it's not like you know,

[19:34] we know, we've talked about,

[19:35] there's many big challenges

[19:38] And I often say I would be

[19:41] if I didn't know something

[19:43] as revolutionary as AI

[19:45] to help with those other challenges.

[19:47] Of course, it's also a

[19:50] But at least, it's one of these challenges

[19:52] that can actually help with

[19:54] - Well, I hope your optimism

[19:59] Thank you so much.

[20:00] Thank you.

[20:01] [upbeat music]