Latest / The Fourcast / UK unemployment soars: is AI already taking our jobs?

Transcript

UK unemployment soars: is AI already taking our jobs? · The Fourcast

0:00
Hello and welcome to the forecast.
0:01
It's not just Tom Cruise and Brad Pitt facing a future on the
0:05
unemployment line. In the week when a convincingly
0:07
lifelike AI video of the two mega stars slugging it out went
0:11
viral and caused a meltdown in Hollywood.
0:13
Unemployment stats here have hit a five year high, with young
0:17
people the biggest losers. So is it simply the result of
0:20
weak growth and higher business costs or are we already seeing
0:23
the impact of artificial intelligence on hiring, whether
0:27
you're an entry level student, white collar worker or Hollywood
0:30
heavyweight? To discuss this, I'm joined by
0:33
our economics correspondent Helia Ebrahimi and Andrea Miotti
0:36
from Control AI, which is a campaign group warning about the
0:42
the dangerous future of AI Helia.
0:45
First of all, how about are the figures and what's causing them?
0:48
Today's figures in the UK you had unemployment up at 5.2% and
0:55
youth unemployment, you talk about decade high at the moment
0:58
16.1%. Most of that is still we're not
1:02
seeing the evidence of AI being the cause for that.
1:06
The biggest cause for that is because entry level jobs go to
1:11
young people and those are the ones that are on the frontline
1:15
when it comes to higher business cups.
1:17
When the chancellor raises taxes, raises National
1:21
Insurance, those are the jobs. That's where companies are not
1:24
hiring. But don't companies need entry
1:27
level people? I mean, why are they deciding
1:30
that they don't need as many people?
1:32
But this, I I'm having this conversation so, so often these
1:36
days with city executives at banks, at professional firms,
1:41
because they are the ones that are now deciding they they don't
1:45
like to talk about jobs disappearing.
1:47
They like to talk about tasks disappearing and hours
1:50
disappearing. But of course it means jobs.
1:52
It means people, what they don't know is if we put in, get rid of
1:57
our graduate recruitment or half graduate recruitment, what
2:00
happens in 10 years? Who becomes the level, you know,
2:04
one manager who becomes the board in 20 years if all of your
2:10
graduate intake. And I think we're still at the
2:12
point where companies don't quite know what to do with this
2:17
big change that we're having. In terms of the impact of AI,
2:22
you think this is the beginning of the storm?
2:25
So I think definitely, you know, it's really difficult to
2:27
attribute exactly which jobs are being lost due to AI, which jobs
2:31
are being lost due to, you know, increased cost of business.
2:34
We cannot really pinpoint that. But what we can see is what the
2:37
intentions of the AI companies are.
2:39
They are very explicit at what their goal is.
2:42
They all, you know, Open AI, the makers of CHA, GPT, XAI, Elon
2:46
Musk's company Anthropic and more.
2:49
They all have as a mission to build what they call super
2:52
intelligence. This means AI that is more
2:55
competent and can replace people at all tasks and can eventually
2:59
out compete humans you know. Are they there yet?
3:02
No. Are they trying the hardest to
3:05
get there as fast as possible? Yes, and.
3:08
You've got the boss of Anthropic, don't you, Dario
3:10
Mundi saying that half of all entry level jobs in an office
3:16
are going to disappear. Half.
3:18
Exactly in terms of what? What is being done already that
3:22
is leading to a reduction in the hiring of white collar jobs?
3:27
Not not the bars in the retail, but what kind of functions is AI
3:33
performing now that means they don't need to hire as many
3:37
graduates? So right now with AI, we're
3:40
seeing the biggest advances on software development.
3:43
This is really a big thing that these AI companies are focusing
3:46
on. Their idea is if we automate the
3:49
development of software, and especially if we automate the
3:51
development of AI, we will be able to go faster and faster and
3:55
faster. So.
3:55
AI will build itself. Exactly.
3:57
This is their explicit goal. How quickly this will happen is
4:00
hard to tell. If we could predict the future
4:02
exactly, we will all be billionaires, but they are
4:05
definitely investing well, there's.
4:06
Certainly billions lost in the markets last week because the
4:10
markets across in the US and and in the UK kind of had this
4:14
moment of panic about exactly this software companies.
4:17
Is AI going to take over all of these roles and make these
4:23
companies obsolete? But that that's still throwing
4:25
ahead and we'll come back to it. But it but in terms of right
4:28
now, yeah, if you go to a consultancy firm or an
4:31
accountancy firm or a legal firm, any of the jobs that you
4:35
know that bright graduates have traditionally been going after,
4:39
I mean you are saying earlier that that that there is also a
4:43
reduction there. Yeah.
4:44
I mean, you said you had PwC recently.
4:47
They had, they had this ambition just a few years ago to hire
4:51
100,000 new people and they've now said they're not going to
4:55
meet that commitment. And in fact, last year they they
4:58
closed down I think about 5000 positions globally.
5:02
And so you've got these mega firms that were creating, you
5:06
know, huge number of jobs now saying, well, we're not sure we
5:11
can do that. I mean, you know, for for the
5:13
last two or three years, while this debate has been running
5:15
hot. Yeah, all the advocates of AI
5:19
have been saying, oh, look, you know, people have been warning
5:22
about the Internet and the television and everything for,
5:25
you know, for years. And technology always gives us
5:28
new tools. They become tools that people
5:30
will use. And we shouldn't panic.
5:33
Is it already here before, Before we knew?
5:36
I think it's starting and you know, a lot of experts have
5:39
been, to be honest, have been warning about this for years.
5:41
It's just this hasn't reached the vast majority of lawmakers
5:45
and the vast majority of governments.
5:46
But and, you know, to to throw a historical analogy, like for
5:50
example, the car did lead to the disappearance of the vast
5:54
majority of horses. You know, like in this case, we
5:56
are the horses, you know, and the horses were headed to the to
6:00
the glue factory after cars were invented.
6:02
So this is what a lot of experts are warning about like Nobel
6:05
Prize winners like Geoffrey Hinton, sometimes even the CE OS
6:08
of these companies themselves. These companies often have a
6:11
kind of a doublespeak where their their lobbyists and spin
6:14
doctors say there's nothing to worry about.
6:17
You know, keep calm, carry on, let us do what we want.
6:20
No regulation while their own CE OS go on the record talking
6:23
about white collar bloodbath or even up to human extinction.
6:27
How? Does it spill over then into the
6:29
blue collar jobs and the and the, and the service industry
6:31
jobs? That's hellier has just been
6:33
talking about in in today's figures.
6:36
I mean, you know, we've all been through the airport and seen
6:39
that funny robot on wheels kind of trundling down the the sort
6:43
of the the restaurants and it's not quite clear what it does,
6:48
but it seems to take away trays. But presumably that was a person
6:53
taking away trays before. I mean, it it, you know, again,
6:58
people people have been saying, well, this is this is now where
7:00
everyone should become a plumber or open a bar or, you know, do
7:03
do a sort of a people to people job.
7:07
But actually, can AI and robotics do some of that too?
7:11
Yeah. So I think 1 interesting thing
7:13
is that 10 years ago, a lot of people even in the AI field
7:16
would say, well, obviously I will be fantastic very quickly
7:20
at formal methods of mathematics, at writing code,
7:24
but it will struggle with human comprehension and understanding
7:28
emotions and new ones. And then we started having,
7:31
first with GPT 3, then with a big explosion of ChatGPT, and
7:35
now with the more advanced models, models that are very
7:38
clearly very capable of understanding new ones.
7:40
Some people are developing far as social relationships with
7:43
these AI systems. You know, they are not exactly
7:46
in love. They're falling in love with
7:47
these AI systems. They are becoming obsessed with
7:50
these AI systems. This is a concerning development
7:52
in society. But fundamentally, this means
7:54
that these AI systems can speak like humans can.
7:57
They can impersonate humans very easily.
8:00
And they're handling a lot of these new ones pretty well, you
8:03
know, against all expectations, while also being really
8:06
competent at maths, really competent at writing code and so
8:10
on. I mean, if if you look at the
8:12
Kung Fu robots from China that we've just seen, it's it's kind
8:20
of astonishing and you watch. It now they can dance, they're
8:23
going to take over our jobs for. Sure.
8:25
Well, I mean, they're pretty, they're pretty sophisticated.
8:29
And, and obviously what what that makes you think is not
8:33
necessarily the bar job or the masseuse, but certainly the
8:37
soldier. Yes, I mean, look, I, I, I think
8:43
the issue is that the the AI, the impact of AI is, is that
8:48
happening so rapidly that people don't know how to keep up with
8:52
it. But I would say that when I
8:55
speak to people within the biotech world, for example,
8:59
where AI is being used quite, quite extensively, or within the
9:04
legal world where AI is being used and is being talked about
9:07
as replacing lots of lowers, there are still problems.
9:11
This is not a system that has yet, he said.
9:14
Is it here now? Certainly when I talk to people,
9:17
you've got, yes, you've got a whole world in California that
9:20
are going to preach that it's going to there's going to be
9:23
super intelligence in three, 4-5 years time.
9:26
But I think in the real world, in terms of integrating AI into
9:31
companies effectively where things things at the high end
9:36
are working, we're still not quite there.
9:38
But if you. Take the law, you know, where
9:41
lots of lawyers will tell you, well, you know, we're still
9:44
going to need human lawyers. We're still going to need human
9:46
judgement and a human interface for clients.
9:50
So, you know, senior lawyers are still going to exist.
9:53
Where do they think senior lawyers are going to come from?
9:56
If. They don't need junior lawyers
9:57
anymore. Exactly right.
9:59
And I think that that's the that's the thing that I think
10:02
boardrooms are struggling with in terms of when, when you talk
10:06
to people at board level who work in professional services,
10:10
what they're talking about is if we go down this route, what
10:16
happens in a decade, in two decades time when we need to
10:20
have managers, there won't be any managers.
10:23
Down this route, aren't we already there?
10:26
I think we've gone down it. I think there is a choice.
10:28
I think there is always a choice and a lot of the narrative from
10:32
the AI industry is to tell us that we don't have a choice.
10:34
But we almost always have. Like also with other major
10:38
technological developments in the past.
10:39
Like you know, we most countries around the world don't have
10:42
nukes and we have nuclear non proliferation for a reason.
10:46
There was a different path where every country has nuclear
10:49
weapons and we are in, we would be in a very unstable world,
10:52
probably we wouldn't even be here recurring this podcast
10:55
today. Instead we chose the path and
10:57
you know, our leaders chose the path where we are restricted
11:01
nuclear proliferation and we have not had a nuclear war since
11:05
World War 2, which is a fantastic achievement.
11:07
So I think we do have a choice. This doesn't mean rejecting all
11:11
AI. I think there is a lot of AI
11:13
systems that are tools that help people do their job better that
11:17
are very useful. You know, things like Alpha Fold
11:20
from DeepMind based here in the UK, It's not only trained on
11:24
data about proteins and it solves scientific questions
11:26
about proteins. That's really helpful.
11:29
What we should avoid is super intelligence AI that is built to
11:33
replace all humans at all tasks and to make us not be in control
11:38
of our economy anymore. And what happens if we don't?
11:41
If we let this run riot? Or to those countries who let it
11:43
run riot because the world won't agree on this.
11:46
So unfortunately, top AI experts, Nobel Prize winners,
11:51
and even the CEOs of AI companies warn that super
11:54
intelligence could cause human extinction.
11:56
This is beyond just jobless. This is literally the end of the
12:02
human species on the planet. And these warnings are very
12:05
serious. And if we continue on this path,
12:08
this is what, again, many top experts warn about.
12:11
And I think we have a choice to not do this.
12:13
We have made this the right choice before in history.
12:15
With dangerous technologies, we can do it again.
12:18
But how do we get people to take that seriously?
12:20
Because the rise of the machines has been a Hollywood theme for
12:23
decades now. And we just kind of think, well,
12:26
that's, that's fictional. I think things are changing now.
12:30
So I am seeing, well, I think just yesterday or two days ago,
12:33
Hugh Grant was talking about the threat from AI and human
12:37
extinction. And this is reaching millions of
12:38
people right now. People like Geoffrey Hinton,
12:41
Nobel, you know, Nobel Prize winner, great British scientist
12:44
has been. One of the godfathers of AI.
12:46
Yes, or the godfather of AI has been warning but.
12:48
But how are they saying it will happen?
12:51
Yeah. So, you know, it's difficult to
12:53
predict exactly the way, but the direction of travel is quite
12:56
clear. If we have AI systems that are
12:59
fully autonomous, vastly smarter than us, integrated, you know,
13:03
across the economy or critical parts of the economy, we have
13:07
essentially handed over control over our world and our economy.
13:12
Our power won't there. Always be an off switch.
13:15
Right now we don't have these off switches.
13:17
I think if we had them, that would be a much better
13:19
situation. But even right now, we don't
13:22
really have them. You know, even if the CEO of one
13:24
of these major companies is concerned and things we need to
13:27
shut it down. This is going too far.
13:29
What is he that he doesn't have a big red button on his desk.
13:32
He needs to call his engineers and his engineers will call a
13:37
data center and they will say, sorry, it's too expensive to
13:40
shut down. It's going to take us, you know,
13:42
a few days to turn it on again, try again next time.
13:44
Like we don't even have the infrastructure to really make
13:48
this choice. I believe we should build it.
13:49
I believe we should keep the choice and the option available
13:53
to turn it off if it's going too far.
13:55
But we don't have it right now and it's time to build it.
13:58
I. Mean are you really saying that
14:01
your group exists to warn against human extinction?
14:04
Yes. Absolutely yes.
14:06
And this is also what top AI experts, Noble Prize winners,
14:09
and even the CE OS of some of these companies are warning
14:12
about. And this is a serious warning.
14:15
And and how seriously do you think the authorities are taking
14:19
it? I think right now most
14:21
governments are doing frankly like nearly nothing, but this is
14:24
mostly because they are not aware.
14:26
So I mean, do they think you're crackers so far when you when
14:29
you put these warnings out? In practice, no.
14:32
So so far, most of them just have never heard about, first of
14:35
all, what AI companies are building.
14:36
They're not just building chat bots, but building super
14:39
intelligence. Again, when it will come is a
14:41
different question, but that's the direction of travel.
14:43
And second of all, in our experience, we started at the
14:46
end of 2024 to just brief all lawmakers in this country about
14:51
the issue, what's a state with AI, where it's going, what the
14:53
experts are warning about. And now, one year later, after
14:56
over 150 briefings, we have more than 100 lawmakers in the UK
15:00
across political parties recognizing the risk of
15:03
extension from AI and calling for binding regulation.
15:05
Can you just explain what what is the difference between asking
15:09
ChatGPT to write a piece of code and super intelligence?
15:13
Yeah, So the ChatGPT you're using right now is mostly on
15:18
your end. It's mostly a chat bot.
15:19
It's a interface where you can ask questions and you get some
15:23
answers already. The ChatGPT that Open AI has in
15:27
their own servers is much more powerful and can do many more
15:31
things. Many of these models are not yet
15:33
released to the public. It can write code.
15:36
It can use computers directly. We have seen this recently with
15:39
the mult book and mult bot situation.
15:43
It can use most tools that you could use from a computer, which
15:47
is a lot of tools in the modern economy and they are making this
15:51
better and better. Super intelligence would be
15:53
that, but smarter than people with access to computers.
15:59
Access to physical infrastructure and much, much
16:03
more capable than the systems that we have right now.
16:06
Heli, I mean, do you hear this as a concern?
16:10
I do in business. I, I, I think, I, I think I
16:14
bizarrely, you hear it most with those within the tech industry.
16:18
It's, you know, you're absolutely right that the
16:21
people. They're the only ones who know
16:22
what's. Possible the the people they're
16:24
they're also within that bubble. So there's a very strange thing
16:28
where tech CEOs have no idea how long it takes to implement
16:33
technology in real society and that that lag sometimes maybe
16:37
are saving grace. But lots of people within
16:40
technology are what warning about these kind of problems.
16:44
It's where when, you know, I think that the problems you have
16:50
is, for example, Europe was very hard on regulation after there
16:54
was a letter saying there could be big problems with AI and
16:58
super intelligence and extinction.
17:00
I think was part of that, that big letter that Europe has
17:05
fallen behind when it comes to technology. the US is completely
17:09
myopically fixated or not falling behind China, who is
17:13
integrated AI and robotics much more effectively across the
17:18
economy than even the US has. So because of these competing
17:24
forces, I think you've got company bosses who go, yeah, I
17:29
see this is risky. I see there's a problem.
17:32
I see the dangers, but we cannot afford to slow down.
17:38
We cannot afford to stop and I think that's the danger.
17:41
But that doesn't mean that people in the UK cannot make
17:46
different decisions. And by the way, there are lots
17:48
of much more prosaic problems that are going to impact the
17:52
jobs market and the economy and be problems created by I.
17:57
Mean, I was going to say. For extinction, so.
18:00
So what? So what are the challenges that
18:01
the Treasury and IT needs to be thinking about and the sort of
18:06
the policy levers that it needs to start thinking differently
18:09
about given the jobs market is going to change dramatically?
18:12
Well, I mean, you need to have some really bold thinking.
18:15
I think you've also got this problem.
18:17
Don't forget Krish, in the UK we have a problem in addition to
18:21
all of this with inactivity. We've got almost a million
18:25
people in the UK who are not in education or training and we
18:30
don't know why exactly but but. Might we not start thinking
18:34
about that differently in a, in a world where you know what,
18:38
what you're both really saying is a lot of people aren't going
18:40
to be inactive and we've got X million people who are
18:44
economically inactive at the moment.
18:47
And and that's just kind of going to be the reality in the
18:50
future anyway. I mean, what if you were Elon
18:52
Musk, you'd say, well, don't worry, don't worry exactly,
18:55
because no one needs a pension, no one needs to work because the
18:58
the bounty from AI, because all of these companies are going to
19:03
be so much more productive now they've got rid of the horses.
19:06
Yeah, that we're all going to be living in, I mean.
19:09
Is it, is it worth it having a, you know, a, a, a new, a new new
19:14
deal trying to get young people into basic jobs that may not
19:21
exist in five or ten years? Time.
19:22
I think you're right to call that out, but the problem you
19:25
have is that when young people do not get jobs early in their
19:29
career, it normally means that they suffer from problems with
19:35
their careers going all the way through their lives and their
19:39
earnings are capped now. Yeah, OK.
19:42
If the world of work completely changes and revolutionizes and
19:46
in a way that you couldn't have anticipated, yeah, you ever you
19:49
have to throw everything out. But I don't know if anyone has,
19:53
you know, in Treasury is thinking about universal basic
19:56
income. You know that that's not on the
19:58
table. We can't even address the
20:01
problem that we can see, let alone a problem that we have no
20:04
sight of yet. What is the argument that the AI
20:08
evangelist make as to how this will produce rapid economic
20:13
growth? So some of them simply say that
20:18
we should build this AI systems and let them take over the
20:22
economy and hope that they are benevolent to us.
20:25
And you know, this might sound like a joke, but they they do,
20:29
you know, it's, it's hard to give the the best case for any
20:32
scenario. Others will generally will say
20:36
that the vague point about all technologies will destroy some
20:41
old activities and create new ones.
20:43
But I think with most of them, especially with most of the,
20:45
the, the CEOs and the executives, less so the, the
20:48
spokespeople, they are quite honest about the fact that
20:51
they're trying to replace most people at most tasks.
20:54
And they're just then they give a token concession about perhaps
20:58
there will be a universal basic income.
21:00
So you know, all jobs will will be gone, but people will be paid
21:04
by the government or by this company that they control or
21:09
simply that we will live in a post scarcity society where AIS
21:12
are doing everything and they're controlling the entirety of the
21:15
economy. And we are a bit like animals in
21:17
a zoo, but we are well fed because AIS keep us around.
21:21
This is what now Elon Musk has moved on to saying he's just
21:25
talking. He's not talking about humanity
21:27
keeping control. He's talking about, well,
21:28
obviously once AI is smart enough, we will just be pets to
21:32
the AI. And let's just hope that we
21:35
don't get put down by the new AI overlords.
21:38
And I think fundamentally, all of these visions are deeply
21:41
unappealing, I think for most human beings.
21:43
You know, I don't want to be a pet.
21:45
I think you all don't want to be pets of AIS either.
21:48
And but I do think we can chart a different path.
21:50
Doesn't mean stopping all AI, but it means drawing a clear red
21:54
line on things like super intelligence that are simply
21:57
very very bad for our species and our society.
22:00
And how do you think it would happen?
22:02
I mean, would it would it have to be a sort of AUN type
22:06
agreement across the world that people agree not to do certain
22:10
things? I think it should be both at a
22:13
domestic and international level.
22:14
I think the. But it hasn't worked, has it?
22:17
Because we had an AI summit where there was supposed to be a
22:20
group that was led by Yoshua Bengio that were going to tie
22:26
people's hands and make sure they believe worked ethically,
22:30
hasn't worked, hasn't proved effective because of all of
22:32
those competing forces. I think frankly, governments
22:34
haven't even tried like, you know, the the group of Yoshua
22:37
Benja is just writing a report. You know, I'm sure, you know,
22:39
Yoshua is doing heroic work in writing this report, but he
22:42
doesn't have powers to intervene.
22:44
And I think, for example, the UK, the Labour government
22:48
promised in its manifesto that they would regulate the most
22:51
powerful, the AI systems. So far we have seen nothing
22:55
being done about as we've seen this commitment dropped but.
22:59
Because of fear basically, again, that the UK economy is
23:03
weak and therefore we have to do everything we can to attract
23:07
growth, even if that is, you know, making a pact with the
23:10
devil. It is a pact with the devil, but
23:12
I think there is a narrow path where we can succeed, where we
23:15
keep the most of the AI systems that are helping us with our
23:19
work. But we say a clear no to super
23:22
intelligence that just. From the idea of regulation, how
23:24
could you regulate AI without AI?
23:28
We, I think we've done a great or pretty good job at least as
23:31
humans regulating a lot of technologies, you know,
23:34
ultimately. Well, I mean that, you know,
23:36
arguably the regulators are struggling with television and
23:39
that's the 20th century technology.
23:41
I mean, when it comes to social media and AI, forget it.
23:44
They're not doing anything. I think it's true that
23:47
institutions have gotten worse and, you know, there's been
23:50
decay, but I think it's also because people are engaging with
23:54
them less, and a lot of governments have started to
23:57
become kind of, frankly, more cowardly about their position.
24:01
You know, governments exist to set the rules of the game so
24:05
companies can compete on the market.
24:06
And this has always been the case.
24:08
You know, when tobacco companies, it was found out that
24:11
tobacco causes cancer, governments stepped in.
24:13
You know, of course, tobacco companies fought tooth and nail.
24:16
They did propaganda campaigns to smear scientists revealing this,
24:20
even though they were well aware about this.
24:21
But governments stood up for their citizens and set the rules
24:25
of the game. This doesn't mean that these
24:26
companies are gone. You know, we still have tobacco
24:28
companies that are pretty. They have pretty good
24:30
valuations. They just were forced to move
24:32
away from the thing that was clearly dangerous to most people
24:35
on to less dangerous and more beneficial economic activities,
24:39
and this can be done again. But listen, you know, if we say,
24:43
you know, before super intelligence and human
24:46
extinction happens, I think there the argument that the tech
24:50
bosses make is there's a whole cache of jobs that we don't know
24:54
about that are going to emerge. We can't see them yet because
24:58
they're going to come for the generation after.
25:01
So you've got this one generation that are going to be
25:04
stuck because essentially they're going to come at the
25:06
time when jobs are disappearing and we haven't yet realized what
25:11
the next possibilities are. But that there will be new
25:15
opportunities based around a more optimistic reading of what
25:20
AI can do in terms of problem solving, in terms of fixing
25:24
things, in terms of creating, you know, a world where scarcity
25:28
is is not as problematic. I.
25:32
Mean that that's what the evangelists say.
25:34
But on on what basis are you repeating it?
25:38
But you say there will be a future.
25:40
Maybe this is the failure of mankind that we can't quite
25:44
picture, you know, baddies in an office trying to, you know,
25:48
allow the world to completely dissolve and fall apart.
25:52
I can't believe that we will get to the brink and not look back.
25:57
I think the problem becomes with technology like AI that you get
26:02
to the brink and you think, no, I want to step back, but it's
26:05
too late because the technology, and I think that's the thing
26:08
about super intelligence is it's no longer the program is no
26:12
longer taking cues from the person.
26:15
Well, what about the fact, Andrea, that, you know, that we
26:19
do actually still first for human contact and human thought
26:24
and human emotion. You know, we could put out a
26:28
podcast created by AI. You know, the technology exists
26:32
now instead of bothering to take up 40 minutes of heaviest time
26:37
when she's trying to get a piece on the news, we could have just
26:41
asked Google Notebook to to make as a podcast about this issue
26:47
with the with these positions. And I, and I do that sometimes
26:51
with research, just our curiosity, say like, make me a
26:54
podcast on this issue. And and it's awful is the truth.
26:57
And you know, you can't bear to listen to it because it doesn't
27:01
have a human quality to it. I mean, is there not some hope
27:06
that just because it can doesn't necessarily mean we will want
27:10
it? I think this is where we need to
27:14
make our choices though, because.
27:15
You know, the same applies to Hollywood and drama and TV and
27:18
all those sorts of things. But.
27:20
I think definitely we do have a choice and like we can choose to
27:22
prefer the human contact and the human interaction, but we have
27:26
to actively make this choice. And if we just if we don't take
27:29
a stand, what's happening is that in practice, more and more
27:32
content is now AI generated. Like, yes, this podcast is not.
27:36
And I'm, you know, I'm glad to be here as a real human being
27:39
and not an AI. But there, there is more and
27:41
more content that is now being produced with AI.
27:43
And we'll probably go into a sort of a hybrid world where
27:46
quite a lot of the movie is AI. Yeah, and it maybe still has
27:49
real. But to your point that it's
27:50
terrible, some of this stuff, you know, I'd be more worried
27:53
that I mean, in Jeff Hinton's list of professions that were
27:58
going to be hit by AI, journalism is front and centre.
28:01
But I'd be more worried if my Netflix AI program could produce
28:06
some movies that I actually want to watch.
28:08
It can't even predict what movies I would like to watch.
28:13
So I don't. I feel like all of, as you said,
28:16
it's not just that we hanker. Maybe that's.
28:18
Because your husband just goes on and watch stuff that you
28:20
don't really like when you're not.
28:21
There it's always possible, it's always human error in the centre
28:25
of it. But imagine, you know, I don't
28:27
think it's just that we hanker after the human touch.
28:32
I think at the moment I can only talk to what I hear at the
28:35
moment. AI programs are useful and they
28:39
are helping with some tasks, but the biggest risks I hear about
28:43
are that they are making mistakes and that they are being
28:46
adopted too quickly. And so people are assuming in
28:52
very important fields that this this program is going to deliver
28:57
the right results and the right research.
28:59
And I've heard that from a lawyer and a scientist recently
29:03
and, and, and that is that is a bigger worry for me in 2026.
29:08
Final thought from you both, if you were 20 years old trying to
29:13
work out what to do with yourself, what should you be
29:16
thinking about Andrea? That's a great question.
29:19
I think, I think ultimately I, I don't have good job advice, but
29:25
one piece of advice is if you find this concerning, if you
29:28
find where this is going concerning contact, you know,
29:30
you're a lawmaker, contact you're politicians, they don't
29:33
bite. They are real people like you
29:35
for now, you know, and they care about what you care about.
29:39
And they often they more and more are hearing from
29:41
constituents that constituents are concerned about AI and
29:44
concerned where AI is going, including the warnings about
29:47
super intelligence. And they want them to ask.
29:48
And if they, if they hear this from enough people, things will
29:52
change. I think we should not be
29:53
nihilistic. I think we can pull back from
29:55
the brink before we get there. I think, you know, what would I
30:00
do if I was 20? I mean, not not become a
30:04
software engineer, that's for sure.
30:06
But I would tell you that today when we we're doing a piece on
30:09
unemployment and we went up to Birmingham and we're talking to
30:13
these young women who have been out of work and really struggle
30:17
to get back into the employment market.
30:20
And one of them who was a massive firecracker, just
30:22
complete pizzazz. And her boss was saying she's a
30:26
force of nature. She's already been promoted.
30:29
She was out of work for four years and that cycle of not
30:33
being in work wasn't just challenging for her mental
30:38
health because she couldn't afford to do all the things she
30:41
wanted and she was struggling with her bills.
30:44
It hit her confidence. So this world where we no longer
30:49
need to improve ourselves intellectually and also we don't
30:55
need to work and we don't need to have purpose, I find it very
30:59
hard to buy into that because I I don't see how people can be
31:04
happy in that world. Thank you both very much indeed.
31:07
That's the forecast on that cheery nose.
31:10
Bye bye.