Latest / The Best of the Money Show / Is AI really coming for your job?
Transcript
- 0:00Speaker 1: Amid all of the concern, and I think deep worry
- 0:03Speaker 1: might be one of those things that keeps you awake
- 0:04Speaker 1: at night about the impact of artificial intelligence on your job,
- 0:08Speaker 1: a really nuanced investigation into this by the Bureau of
- 0:11Speaker 1: Economic Research. And I suppose the top line might be
- 0:16Speaker 1: that while AI may not necessarily take your job away immediately,
- 0:20Speaker 1: it might prevent the economy from creating more jobs, and
- 0:23Speaker 1: particularly for people entering the job market, that really important
- 0:26Speaker 1: first job. Now, Before we go any further with this conversation,
- 0:29Speaker 1: I must just point out it's moving incredibly quickly. If
- 0:33Speaker 1: you're listening to it live, it's kind of up to
- 0:35Speaker 1: the minute. If you're listening to it on tomorrow's podcast,
- 0:38Speaker 1: the situation might already have changed. Corbis Fenter is the
- 0:41Speaker 1: person who's put this together, the senior economist at the
- 0:44Speaker 1: Bureau of Economic Research. Corbis, good evening. I was really
- 0:48Speaker 1: struck by one of your first points, and I'd never
- 0:51Speaker 1: really thought about it, that in countries where AI is
- 0:53Speaker 1: pretty widely used, I think you pointed to Denmark, It's
- 0:56Speaker 1: not actually replacing human beings immediately. How would you sort
- 1:00Speaker 1: of sum up what you believe to be the impact
- 1:03Speaker 1: of AI in the workplace so far?
- 1:06Speaker 2: Well, you know, one of the things we try and
- 1:09Speaker 2: do at the Bureau is be data-led. You know, it's
- 1:12Speaker 2: easy to fall into the hype of it's either wonderful
- 1:15Speaker 2: or terrible. And when you look at the data, it
- 1:18Speaker 2: doesn't really matter to date whether you're using a lot
- 1:23Speaker 2: of AI or or whether you're using no AI or
- 1:26Speaker 2: if you're only relying on it, it hasn't really shown
- 1:29Speaker 2: up in the numbers. What is happening is the nature
- 1:32Speaker 2: of work is changing. So the people that lean into
- 1:34Speaker 2: it become more productive. Over time, I suppose those companies
- 1:38Speaker 2: will do better than the ones not doing it. But
- 1:42Speaker 2: it hasn't really taken any jobs yet, which was an
- 1:45Speaker 2: interesting finding from the study.
- 1:48Speaker 1: I mean, that doesn't mean it won't in the future,
- 1:50Speaker 1: I suppose. And this is really the crux of the problem.
- 1:53Speaker 2: Yes, I think that's the problem is that especially kind
- 1:57Speaker 2: of clerical work and things that rely on existing requirements
- 2:02Speaker 2: might be automated to some extent. And you can see
- 2:05Speaker 2: that starting to happen with fewer positions for law clerks
- 2:10Speaker 2: and article clerks and things like that already starting to happen.
- 2:15Speaker 2: And it might also lead to a shortage of experience
- 2:19Speaker 2: later on because I think one of the dangers that
- 2:21Speaker 2: one can face is to believe the hype and think
- 2:24Speaker 2: that the jobs will stay the same. I think the
- 2:26Speaker 2: whole point of AI or superintelligence, as President Trump has
- 2:32Speaker 2: now termed it, is essentially going to change the jobs.
- 2:37Speaker 2: And the job that we're currently doing won't be the
- 2:40Speaker 2: same in five years' time. And I think that's the
- 2:42Speaker 2: nuance that's harder to read into. And it's nonlinear movements.
- 2:46Speaker 2: It's not necessarily going to be a predictable outcome because
- 2:50Speaker 2: it is a very rapidly evolving technology.
- 2:53Speaker 1: So many different ways to look at it. When we've
- 2:54Speaker 1: seen previous big changes in technology, think of word processing,
- 2:59Speaker 1: for example, we just produced a lot more. I remember
- 3:03Speaker 1: I wouldn't be able to provide the source for this,
- 3:05Speaker 1: but before you could actually type anything, a legal contract
- 3:09Speaker 1: would be a very short document. So you didn't have
- 3:12Speaker 1: many lawyers, one of the reasons you didn't. When typewriters
- 3:15Speaker 1: came and photocopiers, the legal documents and contracts started to
- 3:18Speaker 1: really expand. And the moment you had word processing and
- 3:21Speaker 1: you could send PDFs around, they grew bigger still, and
- 3:24Speaker 1: that meant you actually ended up with more lawyers. You
- 3:27Speaker 1: don't see AI doing that?
- 3:28Speaker 2: I think one of the biggest differences with this one
- 3:32Speaker 2: is AI can be used continuously everywhere by everyone at
- 3:36Speaker 2: the same time without it incurring any extra costs. So
- 3:40Speaker 2: the lawyers, you know, you have one AI agent that
- 3:45Speaker 2: can do the work of all the lawyers theoretically now.
- 3:48Speaker 2: So I think the way to look at artificial intelligence
- 3:51Speaker 2: for me, which is a slightly different different view on
- 3:54Speaker 2: it is the easy answers will be answered easily. And
- 3:57Speaker 2: you have to start asking better questions, which is the
- 4:01Speaker 2: optimistic view on AI. So if you think of current jobs,
- 4:07Speaker 2: If they can be automated, they probably will be over time.
- 4:09Speaker 2: It's a question of time and friction and legal requirements,
- 4:12Speaker 2: et cetera, et cetera. But the companies that evolve and
- 4:15Speaker 2: lean into it will hopefully be improving their output and
- 4:19Speaker 2: creating new, better ways of work. And I think South
- 4:22Speaker 2: Africa's challenge is to make sure that we enable that
- 4:26Speaker 2: kind of adoption in South Africa. So we've got a
- 4:29Speaker 2: couple of lessons we have to be very wary of.
- 4:32Speaker 1: I mean, one of the things that really worries me
- 4:34Speaker 1: is that I don't see us innovating in this space.
- 4:36Speaker 1: We're kind of following. And what that means is we're
- 4:39Speaker 1: a bit behind. I don't hold any particular person or
- 4:42Speaker 1: institution responsible for that. It just is. What that also
- 4:46Speaker 1: means is that we're not actually thinking about how we
- 4:49Speaker 1: then get younger people into their first job. And youth
- 4:53Speaker 1: unemployment is the biggest problem we have. Now, China has
- 4:57Speaker 1: a very high rural youth unemployment problem. So do other countries.
- 5:01Speaker 1: No society is going to survive if young people believe
- 5:04Speaker 1: there is no future for them. AI creates a kind
- 5:07Speaker 1: of almost existential problem for a society. I don't want
- 5:09Speaker 1: to be apocalyptic about it, but it does pose serious
- 5:13Speaker 1: questions for governments.
- 5:15Speaker 2: You're absolutely right. So, I mean, the lesson for South
- 5:18Speaker 2: Africa is that we need to be obsessive about creating
- 5:21Speaker 2: enabling environments for first jobs specifically. So we need to
- 5:24Speaker 2: sort out training, kind of the skills development side. I
- 5:28Speaker 2: know it's almost a swear word. But we really need
- 5:32Speaker 2: to worry about that side. And we also need to
- 5:34Speaker 2: be very wary of raising the cost of employment, either
- 5:39Speaker 2: the cost or the complexity, because that just increases the
- 5:43Speaker 2: likelihood of either automation or simply becoming a traded good,
- 5:49Speaker 2: which we import from elsewhere. So it is a dangerous time.
- 5:53Speaker 2: But the other advantage for a country like South Africa, yes,
- 5:56Speaker 2: we might not be the leaders in this, but it
- 5:58Speaker 2: this is a technology that's ubiquitous and that everybody can use.
- 6:01Speaker 2: So we can also leapfrog a lot of that legacy. Um,
- 6:04Speaker 2: I don't want to call it backwardness, but, but, but
- 6:07Speaker 2: just the fact that we're not at the cutting edge
- 6:09Speaker 2: of it, you know, we can all subscribe to the
- 6:12Speaker 2: top model of Claude, um, and have the same advantages as,
- 6:16Speaker 2: you know, as the American engineer.
- 6:20Speaker 1: Um, You also raised something I never thought of, which
- 6:23Speaker 1: is that if we don't employ enough people, and particularly
- 6:27Speaker 1: people in certain areas of work who earn a certain amount,
- 6:29Speaker 1: there will be fiscal implications. There won't be enough tax
- 6:32Speaker 1: being paid, which means every single person, every single social
- 6:37Speaker 1: grant is suddenly affected by this.
- 6:39Speaker 2: Yeah, it is one of the worrying things. The type
- 6:43Speaker 2: of jobs that might be impacted or will probably be impacted,
- 6:46Speaker 2: but again, the findings, none of this is short-term for
- 6:49Speaker 2: kind of the next two, three years, but 2030, things
- 6:52Speaker 2: will start changing without a doubt, will be professional occupations.
- 6:57Speaker 2: And these are the high-income taxpayers in South Africa, and
- 7:00Speaker 2: we are disproportionately dependent on a very narrow tax community,
- 7:07Speaker 2: as you know. So anything that affects... the kind of
- 7:11Speaker 2: professional levels can have major fiscal implications and that has
- 7:15Speaker 2: major political and social implications. So again, we need to
- 7:18Speaker 2: be wary of that and keep monitoring it and make
- 7:23Speaker 2: sure that we don't do any policy missteps.
- 7:27Speaker 1: Kwebis Fenta, thanks so much. Really appreciate it. Very thoughtful,
- 7:30Speaker 1: very nuanced view on AI, what it's going to mean
- 7:33Speaker 1: for you and your job and for the future as well.
- 7:36Speaker 1: Kwebis Fenta is the Senior Economist at the Bureau for
- 7:39Speaker 1: Economic Research. Look, it's obviously all going to change as
- 7:43Speaker 1: well in many different ways. I see some people who
- 7:46Speaker 1: say there's going to be a lot more jobs in
- 7:48Speaker 1: what they call the authentic economy, face-to-face things, that kind
- 7:52Speaker 1: of stuff. Honestly, I remember very, very clearly some trouble
- 7:58Speaker 1: out of the CEO at Standard Bank. I was asking
- 7:59Speaker 1: him about AI and the number of people they employ.
- 8:02Speaker 1: He thought that AI, they do use it, would probably
- 8:05Speaker 1: make him employ a lot more people in the future.
- 8:07Speaker 1: I think it's Jenkins' paradox. You might recall it better
- 8:10Speaker 1: than I do. I thought that was a very interesting comment.