Introduction and Podcast Background
00:00:02
Speaker
Welcome to HSBC Global Viewpoint, the podcast series that brings together business leaders and industry experts to explore the latest global insights, trends, and opportunities.
00:00:13
Speaker
Make sure you're subscribed to stay up to date with new episodes.
00:00:16
Speaker
Thanks for listening.
00:00:17
Speaker
And now onto today's show.
00:00:24
Speaker
The following podcast was recorded for publication on the 23rd of February by HSBC Global Research.
00:00:30
Speaker
All the disclosures and disclaimers associated with it must be viewed on the link attached to your media player.
Generative AI in Global Economy
00:00:39
Speaker
Hello and welcome to the Macro Viewpoint.
00:00:41
Speaker
I'm Piers Butler in London.
00:00:42
Speaker
And I'm Aline Van Dyne in New York.
00:00:45
Speaker
This week, we're picking up on one of the hottest topics in machine learning, generative AI, and our latest research on the potential it has to affect sectors across the global economy, from banking to pharmaceuticals.
00:01:01
Speaker
And a special guest joins us to explain more.
00:01:04
Speaker
I'm very excited to be here today, even though I don't actually exist.
00:01:09
Speaker
We'll also be diving into the latest economic data from Europe.
00:01:12
Speaker
Is recession avoidable across the region and why?
00:01:18
Speaker
Now from Europe, we head to Brazil to take stock of nearly two months under the administration of Luiz Inácio Lula da Silva, an investor concerned about the fiscal outlook.
00:01:34
Speaker
We're joined in the studio by Mark McDonald, the head of our data science and analytics team, to discuss the new report on generative AI and the rise of the machines.
00:01:44
Speaker
The report looks at the social implications, the impact on economic sectors, and the science behind it.
00:01:50
Speaker
Mark, thanks for joining us today.
00:01:53
Speaker
Thanks for having me on.
00:01:54
Speaker
So we're going to hear from your avatar in a minute.
00:01:57
Speaker
But before that, firstly, give us a quick explanation of the difference between AI that we hear a lot about and generative AI.
00:02:04
Speaker
What is the difference there?
00:02:06
Speaker
So most AI that's been prevalent for the last decade or so is predictive AI.
00:02:12
Speaker
Those are models much like the machine learning models that we include in our research, where you give it some historical data to learn from, if X, then Y, and then you give it new data and it makes a prediction.
00:02:24
Speaker
What's different about generative AI is these are machine learning models that can produce new content.
00:02:31
Speaker
And this is something that, until recently, was viewed as something that, you know, was the unique purview of humans.
00:02:38
Speaker
But it turns out that these machine learning models, they're pretty good at this too.
00:02:43
Speaker
It's something that, you know, has taken the world by storm recently.
00:02:47
Speaker
Indeed, and ChatGPT is now something that everybody is talking about.
00:02:51
Speaker
But is it hype or is there some reality to this?
00:02:54
Speaker
I would say there's reality to this.
00:02:56
Speaker
It's one of these things where sometimes if you speak to people who are on the technology side, they're surprised by the huge hype that there is around this area at the moment because from a technological perspective, things have progressed smoothly.
00:03:12
Speaker
It's been a steady improvement on what was already there.
00:03:16
Speaker
But what that misses is the fact that two years ago, if somebody wanted to use technologies like this, then they had to be someone like me, somebody that could spin up Jupyter Notebooks and program in Python and could go and use these models via code.
00:03:33
Speaker
Whereas now, anybody can just get a login to ChatGPT, and so anybody who can use a website can use these technologies.
00:03:40
Speaker
And it's the proliferation of this technology throughout the economy and getting it out of the hands of geeks like me and into the hands of content producers and entrepreneurs globally, that makes a huge difference.
00:03:53
Speaker
So I think it's not just hype, there is a lot of reality to this.
00:03:56
Speaker
So you talked about the economic implications.
00:03:59
Speaker
Can you give us a few more practical examples?
00:04:03
Speaker
I know in your report you looked at it from a sector perspective.
00:04:05
Speaker
Tell us a bit more about that.
00:04:06
Speaker
Well, yes, we asked our network of analysts from around the world who cover sectors in depth to pull out the implications for their sectors in more detail.
00:04:18
Speaker
So I think some sectors there's clear and obvious immediate implications.
00:04:23
Speaker
So any industry that's involved in producing content, such as the media sector, these models dramatically reduce the cost of producing content and they're already starting to have an influence there.
00:04:35
Speaker
And another sector of the economy where this is already starting to be relevant for investors is in big tech.
00:04:43
Speaker
So we saw a very dramatic example of this with Google when they did a public demo of their model Google Bard, which is a competitor to ChatGPT.
00:04:53
Speaker
The demo was not well received and there was a very sharp reaction in response to this to Google's share price.
00:05:00
Speaker
and clearly as well the semiconductor industry, there's direct relevance there too.
00:05:05
Speaker
However, there are some sectors of the economy where you might not immediately assume that this is relevant, but our analysts think that it is.
00:05:12
Speaker
So the piece has a very sizeable contribution from the healthcare and pharma team, and they go into great depth about the ways in which this technology can be useful to pharmaceutical companies and biotech.
00:05:25
Speaker
So I think there are many areas of the economy where these models are going to start being useful over and above the obvious examples.
00:05:34
Speaker
What about a bit closer to home on the banking side?
00:05:37
Speaker
How does it impact the banking industry?
00:05:38
Speaker
Well, that's an interesting one because obviously we've found this useful in producing this report and the quality of the content that ChatGPT and models like this can produce now is very high.
00:05:49
Speaker
But actually for the banking sector, our banking analysts feel that it will have a pretty muted influence in the banking sector.
00:05:57
Speaker
It's really down to the inherent nature of these generative models.
00:06:02
Speaker
What they're good at is being creative.
00:06:04
Speaker
The last thing you want in a heavily regulated industry is a creative model.
00:06:07
Speaker
You don't want your chat bot for a retail bank making things up when it's dealing with customer requests.
00:06:15
Speaker
And so I think the challenge there is really in getting to the point where the output of these models can be controlled, where you can have guardrails, where you can say, produce me some content, but the content has to refer to these facts in our knowledge base, for example.
00:06:32
Speaker
And then I think it might start to have more of an influence in banking.
00:06:36
Speaker
And that actually brings up a really important point, which are the ethical considerations surrounding this kind of technology.
00:06:43
Speaker
Can you tell us something about that?
00:06:45
Speaker
Well, interestingly, in the video that went with the report, that was one of the questions that we posed to ChatGPT itself.
00:06:53
Speaker
And then our avatar answered the question.
00:06:55
Speaker
So I think it does a better job than I would do.
00:06:59
Speaker
So I'm going to hand over to Artav, my avatar, for that.
00:07:02
Speaker
One major concern is the potential for bias in the data used to train generative models.
00:07:07
Speaker
If the data is not diverse or representative, the model may make decisions that perpetuate existing inequalities.
00:07:14
Speaker
Another concern is the potential for misuse of generative AI, such as creating deepfakes or spreading misinformation.
00:07:22
Speaker
Additionally, as generative AI automates more tasks, there is a risk that it may displace human workers.
00:07:30
Speaker
It is important that we consider these ethical issues and take steps to mitigate them as we continue to develop and use generative AI.
00:07:40
Speaker
Well, that was very clearly presented and convincing.
00:07:44
Speaker
Just to finish on, Mark, can we talk about the sort of broader economic opportunities if this reality really takes off?
00:07:52
Speaker
I think the core influence here is that it pushes the cost of producing content much lower.
00:08:00
Speaker
And once the marginal cost of producing new content goes down, then I think the people who are able to use this new technology appropriately will end up being the winners within a sector.
00:08:12
Speaker
And so, you know, there are always situations when new technology is created where
00:08:18
Speaker
People are worried.
00:08:20
Speaker
But it's not like creating the calculator meant that people stopped being good at maths.
00:08:25
Speaker
And it's not like the creation of these generative AI models will stop people being able to be creative.
00:08:31
Speaker
But probably the people who win and the firms and the sectors of the economy that win will be those that manage to harness this technology and use it.
00:08:42
Speaker
use it effectively to produce content that people still want to see, but to do so more effectively, more efficiently and more rapidly.
00:08:50
Speaker
Mark, Annie did your avatar.
00:08:51
Speaker
Thank you for joining us today.
00:08:52
Speaker
Thank you very much.
Economic Outlook in Europe
00:08:56
Speaker
Now back to the real world, we turn to the latest economic data here in Europe.
00:09:01
Speaker
And for that, we're joined by Simon Wells, our chief European economist.
00:09:05
Speaker
Simon, welcome to the podcast.
00:09:08
Speaker
So, Simon, we've had some economic data early in the week that, dare I say it, would imply that R stands for recovery rather than recession now?
00:09:16
Speaker
Certainly things are turning out a lot better than they might have done.
00:09:20
Speaker
We'd already started to see a bit of a turning point in many of the eurozone indicators, and that was certainly reinforced with the February PMIs.
00:09:30
Speaker
Now, these are widely watched leading indicators of activity.
00:09:35
Speaker
And through the second half of last year, every month they were below 50, which points to falling activity.
00:09:42
Speaker
But in January and February, they've been rising.
00:09:45
Speaker
They are now pointing to rising activity and are consistent with the economy in the eurozone expanding again in the first quarter.
00:09:54
Speaker
Now, if that is the case, it is possible that Europe will avoid recession.
00:10:00
Speaker
And given the size of the inflationary squeeze on real terms incomes, that would be pretty remarkable.
00:10:07
Speaker
I still don't understand what's happened here.
00:10:10
Speaker
You've got to give me a bit more explanation here.
00:10:12
Speaker
Well, you're not alone on that.
00:10:15
Speaker
And one thing we can say is that inflation is high and it is outpacing wage growth.
00:10:20
Speaker
So household incomes are being squeezed and we can say interest rates are rising and that is affecting firms.
00:10:26
Speaker
So what's going on?
00:10:28
Speaker
Well, one possibility is we've had more fiscal support than we thought, and that is keeping people spending.
00:10:36
Speaker
Another possibility is that people are running down some of the savings that they accumulated through the pandemic.
00:10:43
Speaker
But neither of these, I think, are conclusive and not particularly evident in the data.
00:10:50
Speaker
And the other thing is confidence just may be up.
00:10:52
Speaker
In many countries, there were a lot of warnings about how bad the energy crisis might get through the winter through a combination of wind
00:11:00
Speaker
Windy and warmish weather and things just turning out a bit better than expected were actually through the worst.
00:11:07
Speaker
And that may have provided a major boost to confidence.
00:11:11
Speaker
And finally, I'd say is we've seen some revisions to inflation data in Germany.
00:11:16
Speaker
And so it may be that even the income squeeze wasn't as bad as it looked in the data even just a few months ago.
00:11:24
Speaker
So policymakers have been hard at work tightening monetary policy or in plainer terms, increasing interest rates.
00:11:31
Speaker
Are they going to redouble their efforts?
00:11:34
Speaker
And is there a risk that with a lag in terms of monetary policy, they overdo it?
00:11:39
Speaker
Well, that's a very good question.
00:11:40
Speaker
There's a lot of debate about this at the moment.
00:11:43
Speaker
I think one thing that is clear is that all these indicators of stronger activity means that there won't be a deep recession pushing up on unemployment and therefore down
00:11:57
Speaker
You've heard a lot of ECB policymakers coming out over the past few weeks.
00:12:03
Speaker
They do sound pretty hawkish.
00:12:04
Speaker
They are looking at this resilient economy.
00:12:07
Speaker
Even before that, they wondered whether a mild recession would be sufficient to return inflation to target.
00:12:13
Speaker
And therefore, I think in the near term, we are going to see more interest rate rises.
00:12:18
Speaker
So we now expect a 50 basis point rate rise from the ECB in March, which was our existing view.
00:12:23
Speaker
We've now added 50 basis points in May as well, which would take the ECB's policy rate up to 3.75%.
00:12:32
Speaker
Recall it was minus 50 just last summer.
00:12:37
Speaker
So that's a big tightening and you're right.
00:12:39
Speaker
The question now is, is there enough already in the pipeline, given those lags with which monetary policy operates,
00:12:46
Speaker
such that they've done enough.
00:12:48
Speaker
We think maybe there is.
00:12:49
Speaker
The June meeting seems a long way off, so perhaps they pause.
00:12:54
Speaker
But for sure, policymakers will be debating these lags and also the risks of doing too much versus doing too little.
00:13:03
Speaker
The fullness of time will reveal.
00:13:06
Speaker
Well, let's see how it turns out.
00:13:07
Speaker
But thank you for joining us today.
Brazil's Fiscal Outlook Under Lula
00:13:14
Speaker
Ana Madeira, our chief Brazilian economist, has been looking at the policies of the nearly two month old Lula administration, and she's here to discuss her findings.
00:13:25
Speaker
Ana, there are concerns among investors about the fiscal outlook.
00:13:30
Speaker
Tell us where things stand at the moment.
00:13:34
Speaker
So indeed, the biggest concern is about the fiscal outlook, especially because we are going to see a shift between the previous government, which was right leaning into this new Lula administration, which is left leaning.
00:13:52
Speaker
And when we look at the initial speeches from all the newly appointed ministers, they have a heavy focus on social support and they point to the return of old policies that we saw during the previous administrations of Lula and Dilma.
00:14:10
Speaker
And we know that...
00:14:13
Speaker
especially during these previous policies that were used, this led to a very big fiscal burden.
00:14:22
Speaker
So the question for this year is, with all that pressure to do more fiscal spending, is the new fiscal rule that is going to be announced this year going to be fiscally responsible and whether it's going to allow for gross debt to GDP to stabilize in the medium to long term?
00:14:42
Speaker
So this is the big concern for this year.
00:14:46
Speaker
And Anna, in your report, you look at some of the policies that have been announced or signals.
00:14:52
Speaker
Can you talk us through some of the main areas of concern and focus?
00:15:00
Speaker
So we've got a few policies, a few measures that have been announced since January.
00:15:05
Speaker
And we also look in the report at some of the measures that are being signaled.
00:15:12
Speaker
So they haven't been announced, but they are being signaled.
00:15:15
Speaker
This is as important as the ones that have been announced.
00:15:19
Speaker
So among the measures that have been announced, they do have a sense of déjà vu.
00:15:23
Speaker
So, for example, we have the increase in the ministries that President Lula announced at the beginning of this year when he took office.
00:15:33
Speaker
We also have the announcement of the parallelization of all the privatization projects that were ongoing.
00:15:41
Speaker
And as part of the measures that have been hinted,
00:15:46
Speaker
from the new government, we find a bigger role for the public banks.
00:15:51
Speaker
We find a will to change the state-owned company laws and more recently as well, an attempt to rise the inflation target in order to try and reduce interest rates faster.
00:16:08
Speaker
So Anna, do you think all these fiscal changes will be made?
00:16:13
Speaker
What are you actually expecting for this year?
00:16:17
Speaker
That's a good question because timing, in our view, is going to be very important.
00:16:22
Speaker
We think that despite of all these measures that the new government has been signaling, they are going to focus on, they're going to have two main priorities for this year.
00:16:34
Speaker
One of them is going to be to present the new fiscal framework and the second one is going to be the tax reform.
00:16:41
Speaker
And in these two cases, the economic team is hinting at a new fiscal rule that would keep fiscal discipline and at the tax reform that would simplify the tax system, which is positive in our view.
00:16:56
Speaker
And this is relevant because these two measures, these two reforms are very heavy reforms that are going to require several months of discussions in Congress before being approved.
00:17:09
Speaker
And this is the point that is relevant in our view, because this means that it will need, the government will need to postpone eventual market and friendly measures as a consequence of all that effort and all that time being put into the fiscal and the tax reform.
00:17:28
Speaker
And what this means is that as a consequence, it should limit the big shifts in economic policy in the short term.
00:17:40
Speaker
we end up delaying, if you'd like, a little bit the risks for Brazil.
00:17:45
Speaker
So in our view, if we put it into a nutshell, we can say that for us, Brazil, the risks that we see for Brazil are being more medium term in nature rather than short term in nature.
00:18:01
Speaker
Anna, thank you so much for that update.
Conclusion and Future Episodes
00:18:07
Speaker
So that's all from us.
00:18:08
Speaker
If you don't mind, I'll take over and wrap this up.
00:18:12
Speaker
Thank you Piers and Aline.
00:18:14
Speaker
That's all the time the humans here at HSBC Global Research.
00:18:19
Speaker
How for this week's podcast.
00:18:21
Speaker
Special thanks to Mark McDonald, Simon Wells, and Anna Madeira.
00:18:27
Speaker
And of course, to chat GPT and Synthesia.io for giving me a voice.
00:18:33
Speaker
Piers will be off on holiday next week.
00:18:35
Speaker
Who knows, maybe I will be his replacement host from next week alongside Aline Van Dyne.
00:19:00
Speaker
Thank you for joining us at HSBC Global Viewpoint.
00:19:03
Speaker
We hope you enjoyed the discussion.
00:19:06
Speaker
Make sure you're subscribed to stay up to date with new episodes.