Latest / DEEP INSIGHT / Google boss says trillion-dollar AI investment boom has 'elements of irrationality'

Transcript

Google boss says trillion-dollar AI investment boom has 'elements of irrationality' · DEEP INSIGHT

• 0:00
  Speaker 1: Now let's move on, because the head of Google has
• 0:03
  Speaker 1: told the BBC that no company would be immune if
• 0:06
  Speaker 1: the current high valuation of AI tech firms turns out
• 0:09
  Speaker 1: to be a bubble and bursts. In the pitch, I
• 0:12
  Speaker 1: said that the boom in artificial intelligence had what he
• 0:15
  Speaker 1: described as irrational elements. Mister Pitchie, who has run the
• 0:18
  Speaker 1: company for a decade, also said AI models should not
• 0:21
  Speaker 1: be blindly trusted, and more that governments would have to
• 0:25
  Speaker 1: invest in more infrastructure to cope with the technology's immense
• 0:28
  Speaker 1: energy needs. Here's what some of what he told our
• 0:30
  Speaker 1: economics editor pisal Islam.
• 0:33
  Speaker 2: Well, it's an extraordinary moment by even by Silicon Valley standards.
• 0:38
  Speaker 2: You know, every decade or so, you know, you have
• 0:41
  Speaker 2: this inflection points. You know, you have a new technology.
• 0:45
  Speaker 2: It was a personal computer at one point, the Internet
• 0:49
  Speaker 2: coming in the late nineties. Then it was mobile, then
• 0:55
  Speaker 2: it's been cloud what we call as cloud, and now
• 0:58
  Speaker 2: it's clearly the air of AI.
• 1:01
  Speaker 3: And you use the sense of the scale, the sheer
• 1:04
  Speaker 3: amount of investment going in.
• 1:06
  Speaker 2: You know, maybe four years ago Google was spending less
• 1:09
  Speaker 2: than thirty billion dollars per year. This year that number
• 1:13
  Speaker 2: is going to be over ninety billion dollars, and if
• 1:17
  Speaker 2: you collectively add what all the companies are doing, you
• 1:21
  Speaker 2: know we have well over a trillion dollars of investment
• 1:25
  Speaker 2: going in in building the infrastructure for this moment.
• 1:28
  Speaker 3: Now, you mentioned some of those phases of technological advancement
• 1:32
  Speaker 3: that happened with much market excitement as well. And the
• 1:35
  Speaker 3: obvious question, it's around the whole of this country and
• 1:37
  Speaker 3: the whole the world right now. Is is it a bubble?
• 1:39
  Speaker 2: You see real demand and we are constrained in our
• 1:42
  Speaker 2: ability to serve that demand. So you know so, and
• 1:47
  Speaker 2: given the potential of this technology, the excitement is very rational.
• 1:54
  Speaker 2: It's also true when we go through these investment cycles,
• 1:59
  Speaker 2: you know, there are moments we overshoot right collectively as
• 2:02
  Speaker 2: an industry. We can look back at the Internet right now,
• 2:05
  Speaker 2: there was clearly a lot of excess investment, but none
• 2:10
  Speaker 2: of us would question whether the Internet was profound or
• 2:13
  Speaker 2: did drive a lot of impact. It's fundamentally changed how
• 2:15
  Speaker 2: we work digitally as a society. I expect AI to
• 2:20
  Speaker 2: be the same. So I think it's both rational and
• 2:23
  Speaker 2: there are elements of irrationality through a moment like this,
• 2:27
  Speaker 2: no company is going to be immune, including us. If
• 2:30
  Speaker 2: you overinvest, you know we'll have to work through that phase.
• 2:33
  Speaker 2: For many years, we've taken a deeply differentiated approach to
• 2:37
  Speaker 2: how we've approached AI, so we are better positioned to
• 2:40
  Speaker 2: take a long term view and approach this moment.
• 2:43
  Speaker 1: Well, we've me now as a reconomic Islam and our
• 2:46
  Speaker 1: AI correspondent, Mark, thank you both very much for coming
• 2:49
  Speaker 1: in faisal quite an interview. The AI bubble has been
• 2:54
  Speaker 1: talked about for some time. It is significant and really
• 2:57
  Speaker 1: interesting to hear what we heard there from CINDERPITCHAI.
• 3:00
  Speaker 4: So I think it's important to get the context for
• 3:01
  Speaker 4: those people who don't necessarily follow financial markets and sort
• 3:04
  Speaker 4: of going out Silicon Valley to the Mountain View headquarters
• 3:07
  Speaker 4: of Google, you one really got a sense of a
• 3:08
  Speaker 4: total frenzy in terms of this AI investments, and you know,
• 3:12
  Speaker 4: all those things we're typing in, all those prompts, we're
• 3:15
  Speaker 4: typing in about the sheer infrastructure needed, the energy needed,
• 3:19
  Speaker 4: and you're talking trillions of dollars and you just mentioned
• 3:21
  Speaker 4: there they trebled their investment thirty to ninety billion a year.
• 3:25
  Speaker 4: Other companies are doing the same. You talked about trillion
• 3:27
  Speaker 4: dollars of investment going in. So if you're like, that
• 3:29
  Speaker 4: is what we're coming off, and so the question arise,
• 3:31
  Speaker 4: is how much of this is real?
• 3:33
  Speaker 2: Now?
• 3:33
  Speaker 4: His answer to me was that a lot of it's real.
• 3:35
  Speaker 4: It's going to change the world. They would say that,
• 3:37
  Speaker 4: wouldn't they. But I think there's some truth to that.
• 3:40
  Speaker 4: But elements of it reflect a certain amount of rationality.
• 3:43
  Speaker 4: I think of the markets, what you're seeing is a
• 3:44
  Speaker 4: distinction between the big companies that have massive revenues. Google
• 3:47
  Speaker 4: scored one hundred billion dollars of revenue in a quarter
• 3:50
  Speaker 4: from existing search business and some AI others of general
• 3:53
  Speaker 4: Amazon are generating cash and they're spending it almost all
• 3:58
  Speaker 4: on this investment in AI. There are other parts of
• 4:01
  Speaker 4: this ecosystem where they're borrowing the money, slightly weird deals,
• 4:05
  Speaker 4: exotic lending. I think that is probably where the markets
• 4:08
  Speaker 4: are looking, and it's probably where Suchai was suggesting. There
• 4:12
  Speaker 4: were elements of irrationality. That doesn't mean the whole AI
• 4:14
  Speaker 4: thing is going to collapse. Even during the dot com bubble,
• 4:16
  Speaker 4: the famous one twenty five years ago, remember Amazon's share
• 4:20
  Speaker 4: price went down to six dollars. They're now four trillion
• 4:22
  Speaker 4: dollar company. So I think people are waiting for a shakeout,
• 4:25
  Speaker 4: so interesting views from him, and the sort of implicit
• 4:29
  Speaker 4: thing there was that Google is better positioned, in his view,
• 4:32
  Speaker 4: to ride out what might be a shakeout that a
• 4:34
  Speaker 4: lot of people are talking about because they're invest in
• 4:37
  Speaker 4: things like massive super chips as well as the big bubbles.
• 4:39
  Speaker 1: Absolutely, Mark, what was your take from the interview?
• 4:42
  Speaker 5: There's an element of digital Darwinism, if you like. Going
• 4:45
  Speaker 5: on here. Jeff Bezos a couple of months ago sort
• 4:47
  Speaker 5: of alluded to similar things that the strongest companies will survive,
• 4:52
  Speaker 5: and there are sort of you know, it is reminiscent
• 4:54
  Speaker 5: of the dot com bubble at the turn of the century.
• 4:57
  Speaker 5: But a lot of this is about valuation. That Google
• 4:59
  Speaker 5: is in a really really good position in terms of
• 5:02
  Speaker 5: it's very diversified search business is absolutely enormous. It owns
• 5:06
  Speaker 5: companies like YouTube, and then you've got new companies into
• 5:09
  Speaker 5: the AI space like open Ai that have these huge
• 5:12
  Speaker 5: valuations at five hundred billion dollars or something in that
• 5:14
  Speaker 5: in that kind of region or higher than that. Yeah,
• 5:17
  Speaker 5: and they don't have the same kind of product. Yeah,
• 5:19
  Speaker 5: they don't have the same kind of product lineup. They're
• 5:21
  Speaker 5: not they don't have the same kind of revenue that
• 5:23
  Speaker 5: Google has, so Google is very very insulated from some
• 5:26
  Speaker 5: of this talk of a bubble where companies like open
• 5:29
  Speaker 5: Ai certainly aren't stay with for a moment.
• 5:31
  Speaker 1: We're just going to mention that our viewers on BBC
• 5:36
  Speaker 1: two are leaving as we are staying here on the
• 5:38
  Speaker 1: news channel. If you want to follow this discussion, we
• 5:42
  Speaker 1: are here on bus out of the day. I'm getha
• 5:44
  Speaker 1: grom Muthny, sorry to interrupt you, and fires well you
• 5:51
  Speaker 1: might know, and Martin is that I am just new
• 5:53
  Speaker 1: into AI because I'm doing a master's at AI ethics,
• 5:57
  Speaker 1: so I'm trying to get my head around it. But
• 5:59
  Speaker 1: the job Bob's question and the retraining questions, reskilling is
• 6:03
  Speaker 1: one that comes to the forefront of any discussion you
• 6:05
  Speaker 1: have on a What did some of the pitch I
• 6:07
  Speaker 1: teach you about that? Because like, what do we tell
• 6:09
  Speaker 1: our children? What do we do ourselves?
• 6:10
  Speaker 2: What?
• 6:10
  Speaker 4: It's quite striking you see this sheer amount of investment
• 6:13
  Speaker 4: kind of what's this actually going into? And I put
• 6:16
  Speaker 4: it to him, it's about automating many human tasks? Is
• 6:18
  Speaker 4: it not human jobs? Potentially livelihoods? And you know, I
• 6:23
  Speaker 4: think when you talk to some of these tech bosses.
• 6:26
  Speaker 4: They want to stress the win win. Oh, it's okay,
• 6:28
  Speaker 4: everyone can retrain and it's all right, and you know,
• 6:30
  Speaker 4: but like what I do detect now is a certain
• 6:33
  Speaker 4: amount of reality, and I think some of our viewers
• 6:35
  Speaker 4: will be seeing it now if they've got graduate, you
• 6:38
  Speaker 4: know age kids, they'll be seeing it quite tough in
• 6:40
  Speaker 4: the jobs market. Some of the jobs that we would
• 6:42
  Speaker 4: expected in creative industries, in the law, in accountancy, they're
• 6:46
  Speaker 4: proving tougher because some of them are being automated via
• 6:49
  Speaker 4: these large language models, not always successfully in terms of
• 6:53
  Speaker 4: some of the issues around that. But yes, I think
• 6:57
  Speaker 4: he acknowledged that, and of course his answer was you
• 6:59
  Speaker 4: training self up using AI, use the tools and if
• 7:02
  Speaker 4: you know how to use it, you'll be better placed
• 7:04
  Speaker 4: in the jobs market.
• 7:05
  Speaker 1: Well, very quickly, the question of trust and truth was
• 7:07
  Speaker 1: also came up of people just navigating a new frontier
• 7:10
  Speaker 1: where they type in something on AI and they should
• 7:14
  Speaker 1: know that they've got to go and check it with
• 7:15
  Speaker 1: a search.
• 7:16
  Speaker 5: Everybody now has to absolutely question everything they see on
• 7:20
  Speaker 5: social media, for instance, and they also have to question
• 7:22
  Speaker 5: what they do if they use a chatbot to try
• 7:24
  Speaker 5: and determine answers from it. Because they make stuff up,
• 7:27
  Speaker 5: they hallucinate, they get things wrong all of the time.
• 7:29
  Speaker 5: A lot of these products are not ready for primetime
• 7:32
  Speaker 5: quite yet. They have quite a lot of utility, absolutely,
• 7:35
  Speaker 5: but if you're relying on them for accurate answers, for
• 7:38
  Speaker 5: replicable answers, then you're looking at the wrong kind of
• 7:42
  Speaker 5: tech quite yet.