At Livewire’s 2017 investor forum, artificial intelligence (AI) was pegged as the #1 theme that investors need to be aware of. We're already seeing the impacts in our day to day lives and the influence of AI is set to change whole industries and the jobs associated with them.

Just 9% of ASX listed companies currently embrace AI, that’s half the level of our global peers. Undoubtedly, the influence of AI will extend to those companies in listed markets – the question remains as to who will be the winners & the losers. To help Livewire readers understand the universe we’ve invited a panel of experts to share their insights.

Access the content below via video, podcast or read the edited transcript.

The Panel

Lachlan MacGregor, Alphinity Investment Management

Andrew Charlton, AlphaBeta

Niki Scevak, Blackbird Ventures

Interesting quotes and statistics

On market size...

"IDC forecasts that say the AI industry currently is about 12 billion dollars, and it's going to grow to 58 billion dollars."

On AI in 2017...

"It (artificial intelligence) is a lot bigger in the media at the moment than it is in real life, but I think that's about to flip."

"80 percent of the viewing time on Netflix is from things that are served up to them under the recommendation engine rather than from a search."

On displacement of jobs...

"There's also a false choice between human and machine, and I think throughout history, it's the human enhanced by the machine that has always won that equation."

"Every wave of change has generated some displacement and ultimately, humans have found new ways to create jobs and employment."

On adoption of AI...

"I think Australia has generally been quite a long way behind the eight ball on some of these technologies, particularly amongst the larger companies… almost nine percent of Australian listed companies exhibit characteristics of engaging in automation and AI, that's about half some of our global peers."

Topics

1. 0:24 What is artificial intelligence (AI)
2. 3:03 How big is the market?
3. 6:00 Current applications of artificial intelligence
4. 9:45 What’s coming next?
5. 12:53 Which companies are at the forefront?
6. 19:02 What impact will AI have on jobs?
7. 22:30 Which jobs will be most impacted?
8. 24:03 The industries early investors are backing
9. 25:10 ‘Winners & Losers’ on the ASX (TLS, HSO, WOW, BHP, CBA)

Watch...

Listen...

Read...

Welcome to Livewire, today we're going to be talking about artificial intelligence. I'm Lachlan MacGregor, I'm a Portfolio Manager at Alphinity Investment Management. I'm joined today by Niki Scevak, Niki's a co-founder and partner at Blackbird Ventures, and by Andrew Charlton, who is a Director of AlphaBeta.

Today with artificial intelligence everyone talks about this, but it would be great to just narrow in on what we actually mean by artificial intelligence, maybe just in one sentence if each of you could just describe what it means to you. Niki.

Niki: I think artificial intelligence is letting computers spot patterns in information or interpreting images and videos like a human would in classifying objects and so on. I think the definition for me is where a computer can process images and videos like it processes texts today.

Andrew: Artificial intelligence is machines thinking in the way that humans do. Machines going from doing computation or calculation to doing problem solving, and even learning, that's the essence of artificial intelligence.

It's been around for so long, why has it accelerated in the last few years?

Niki: It was a strange coincidence, but people who play computer games have very expensive graphics cards, 10 or 12 years ago, you would buy a computer with a CPU, but you might also upgrade your graphics card. The graphics card contains these special kinds of processes called GPUs, which let computers calculate lots of numbers in parallel to simplify it dramatically. What happened was these graphics cards that were used by people playing computer games turned out to be really good at running artificial intelligence algorithms. And so by virtue of people spending lots of money on graphics cards, those got cheaper. Those got cheaper and cheaper such that artificial intelligence folks started using them, and they got even cheaper and the company that's leading the charge in that respect is Nvidia, which started off for these kind of hardcore computer gamers, but now is forging the way in autonomous vehicles or more general purpose AI processes in data centres or cloud services.

That getting really, really cheap and then just the software world has turned in from a sort of proprietary, think Oracle, IBM world of the 80's, where now lots of software is freely available and open source, and that being so widely available that people can just go and experiment. They don't need permission to start. They don't need money to start, and then they start to make the breakthroughs and again, the world moves forward

How big is the artificial intelligence market?

Lachlan: Many people try and put a size on the industry, or any industry to analyse. I've seen IDC forecasts that say the industry currently is about 12 billion dollars, and it's going to grow to 58 billion. It's anyone's guess, but Andrew, perhaps if you can talk about you think about the size and impact of artificial intelligence on the world.

Andrew: Well 58 billion is still pretty tiny in the scheme of a massive economy, but the role of artificial intelligence is to transform every industry. We're already seeing artificial intelligence having a massive impact in digital industries, but it's also got a huge role in traditional industries as well, becoming a really core part of the way that those businesses operate.

Artificial intelligence, as Niki was saying, is exploding at the moment. We have a really rapid increase in data availability, a really rapid increase in processing ability, particularly parallel processing, plus much better algorithms. So the three of those things are coming together to create an explosion. And how fast it moves will be a function of how fast we adopt it

Lachlan: The whole tech spending globally is about four trillion, so 58 billion does seem like a drop in the ocean compared to that.

Andrew: Yeah, it's a lot bigger in the media at the moment than it is in real life, but I think that's about to flip.

Niki: I think market size is also an interesting area in that where technology companies were happy to sell technology, technology companies have actually become a lot more ambitious and want to be that company.

So whether ... Uber, 20 years ago, would have sold its mobile software to taxi operators all around the world and you would have had the taxi operators spend an X amount of money on the software that Uber might have provided them. Instead, what happened was Uber decided to be the service, to take the responsibility for the consumer experience and to own all of the consumer relationships and to be the business.

Usually, how the world changes is that there's these big waves of change, and when there's a wave of change, it opens up a little window where startups can create big businesses, other technology companies or other companies can use that window to reinvent themselves and to enter new markets and so on and so forth.

And so even perhaps as technology companies have become much more ambitious and become much less likely to say, "Oh, why don't I sell this software to financial services?" Or "Why don't I sell this technology to whatever industry," it's "Let's become an automotive manufacturer, let's become a bank, let's become…" Which, again, the spending there is a little more nuanced because it's a company transformation, it's not our dollars changing hands between a technology company and a traditional company.

Current Applications of Artificial Intelligence

Lachlan: One thing I think that's often underappreciated is that we're already using this technology in our daily life. Maybe if each of you could give an example of something that we're using currently that involves artificial intelligence.

Andrew: We all have artificial intelligence in our pocket. Siri or Siri's Microsoft sister, Cortana, these are artificial intelligence products. The products that enable us to do face recognition, every time we turn on Netflix and get suggestions for what to watch, that's a machine doing problem solving and serving us an answer.

Lachlan: I heard a statistic that Netflix ... 80 percent of the viewing time on Netflix is from things that are served up to them under the recommendation engine rather than from a search. So it's core to that business's proposition.

Andrew: Absolutely. And really interesting to see people's preferences reinforced in a way that in previous generations they might not have been. And what that does to our viewing habits is, I think, really fascinating as well.

Lachlan: And Niki, what's your favourite example?

Niki: It's a little bit not in our daily lives at the moment, but I think the biggest expression of AI will actually be cars driving themselves and what that kind of world that facilitates, whether it is people no longer owning cars, whether ... I think 40 percent of LA is associated with parking, so car parks or parking lanes or driveways and garages, and what do you do with all of that real estate? What do you do with car insurance when the cars don't crash? It just has such a domino effect on how cities work that I think at least in my mind, that will have the biggest impact on our society.

Andrew: I think that example is fascinating. We spent the last 40 years trying to build as many roads as we can to end congestion. That relies on us solving for humans driving cars at 10, 15 metres apart. In an AI world, those cars could be three times closer to each other, and that means you need three times less road area, so in the future we might have the problem of too many rather than too few roads. The consequences of this is just so extensive across a range of different infrastructure, social, health, and education outcomes.

Lachlan: It's interesting how it transforms both our lives and also all of these industries, as you say, autonomous driving will transform not just the automotive industry, it'll transform so many other industries around that. Healthcare, if there's less accidents, less people in hospital. If there's no insurance needed or it's done at a central level, perhaps. Volvo takes out an insurance policy for the whole group.

One of my favourite examples of how this is going to change the world in the future is actually, generative design tools. So it's actually possible now to get a computer to design a new product. You just give it the constraints, you say, "Okay, I want you to optimise for this and that" and get it to iterate, and it will actually come up with a whole new design of something. An aerial drone chassis I saw an example of, and you just give it ... You say, okay, you've got a drone so it's got four propeller spots, you want it to be lightweight, aerodynamic, strong, and see it iterate. And you can actually watch this online and see all the different versions of this. And the final version that came out is just like a squirrel's pelvis. The example being, this is just evolution. This is computers iterating in an evolutionary way to come up with a better design than any human actually could. And these designs look very odd, they're not something we would think naturally.

Transformation and what’s coming next?

Niki you're at the forefront of the next wave of things that are going on. What do you see coming next in terms of transformation?

Niki: I think it's interesting, technology in general has caused massive deflation in the prices of so many industries. You just think about the mobile phone being in your pocket and all of the available services and how much you pay for that per month. But there's been these strands of industries that have not been subject to that technology deflation. Health care being an obvious one, the construction industry's had no productivity gains in 100 years or something like that. And so there's these outlier industries that have kept on high inflation, and I think particularly with AI, it has the ability, obviously in health care, to do the job of a very highly paid person, essentially for free. To diagnose, with greater accuracy, cancers of all types, way earlier than humans are able to detect, and making health care a 10 dollar a month product, not a four or five hundred dollar a month product.

And also, AI allows computers to solve physical problems. So if you think about computers at the moment, it's very information processing heavy, it's virtual problems, it's banking, financial services. It's transformed those industries. Computing hasn't transformed the construction industry, but now, with AI, it essentially allows robots to interact with the physical world, so it allows computers to solve physical problems, rather than just virtual problems.

So I think you look at the economy and you look at the strands of the industries that just have kept on going up in price. I think you'll start to see those moderate and then come way down in price like health care and like construction.

Lachlan: Andrew, you've done a lot of work for some of these big companies. What do you see coming in the future?

Andrew: I think that's right. I think there is a potential to transform industries like health care. But I think there's also a huge opportunity to transform the way we live. To take the health care example, not only can we have artificial intelligence solving health care problems, reading x-rays and other images, you've got the capacity for artificial intelligence to prevent us from getting sick in the first place. The combination of artificial intelligence and the internet of things, tracking our personal data, looking at what's in our fridge, encouraging us to eat better, controlling the steps we do in the day, communicating with health care professionals, tying that back to a range of different types of information from our family and family history, and identifying problems before they emerge rather than after they emerge. That's a pervasive impact on the way that we live and the way that our health care sector will operate, that goes beyond productivity into different models of delivery.

Which companies are at the forefront?

Andrew: The big famous ones are obviously right up there at the forefront. Google's made AI first their mantra, obviously. Facebook is making massive investments in AI and that's really core to their business strategy and operations. But there's also just a pre-Cambrian explosion of smaller companies out there trying interesting things, including many in Australia.

Lachlan: What are some of the interesting smaller companies that you see globally or locally?

Niki: There's a ... We're an investor, but a company called Zoox which was founded by Melbourne CEO who's now living in California, but the company is Zoox and it's reinventing the car from the ground up, so redesigning it to be more of a living room on wheels than something with a driver and a steering wheel and so on. And again, an Uber-like service where you pull out your phone, you hail a Zoox, it takes you to your next destination, and you pay for it one trip at a time. You don't buy a car as such. I think it's incredibly interesting startup.

Obviously, Google, with its own self-driving car effort, and then GM acquired a startup called Cruise. Those are the three credible players in the self-driving car market.

And then I think some of the breakthroughs in diagnosing medical imaging. There are a bunch of early stage startups that have each taken a domain of health care, whether it's a particular type of cancer or a particular type of disease, and using AI to determine, again, really early on, whether someone will have something, not even diagnosing that they have it, but that they will have it in a short period of time.

So I think those are the two areas for me.

Lachlan: Fascinating. You mentioned Google before. Everyone talks about the AlphaGo example, where the computer beat Lee Sedol, who was the world champion Go player. But just recently, they had an AlphaGo Zero, have you heard of that? This attempt? They didn't give it any of the human examples. So typically, you start off with a whole lot of examples of how humans play, whether it's Go or chess or whatever example you're doing, and then you try and train it on that, and then you get it to play itself to become better than humans.

In this example, they had actually started with nothing, they just gave it the rules of the game. In three days, it beat all of the other AlphaGo previous versions, and in 40 days it had developed all of these strategies that had taken 3,000 years for lots of people to come up with how to actually play this game. It's quite incredible, going from being ... Having to train these on lots of data to allowing the computer almost to train itself.

Niki: I think the other incredible thing to think about is AI benefits from all the other AIs before so normally, you learn how to drive your car, you teach your son or daughter how to drive the car, they're starting from scratch and you're trying to pass on your knowledge, versus that AlphaGo algorithm having all of the experience of those such that the next algorithm that's playing a game of Go benefits from all of the knowledge that comes before it. It's just flawlessly passed on to the next iteration of it.

Instead of learning as one driver in the self-driving car, will learn from millions before you who have driven, so that networked, cumulative aspect of it is also very hard to fathom and is very easy to underestimate.

Andrew: I think that'll be really powerful and already becoming really powerful in the call centre space. This Australian company, Digital Workforce, that worked on the development of Nadia, which was the digital call centre interface for the National Disability Insurance Scheme, and learns over time, in a way that no individual call centre operator could learn despite months, years of training, to answer every possible question, to respond to the voice and facial stimulus of the person that they're interacting with, to express moral empathy, recognising the emotion in the person's voice, and have this unlimited bank of knowledge to answer every question.

So I think that space of learning will be really powerful in terms of the interactions with humans as well as problem-solving.

Which ASX listed companies are embracing artificial intelligence?

Lachlan: Are there any listed Australian companies that you think are doing a particularly good job in this space?

Andrew: I think Australia has generally been quite a long way behind the eight ball on some of these technologies, particularly amongst the larger companies. AI more broadly defined as AI in automation, almost nine percent of Australian listed companies exhibit characteristics of engaging in automation and AI, that's about half some of our global peers. We do have some ... Australia's traditional strength industries tend not to be the digital industries.

We tend to have comparative advantage in mining, energy, agriculture. Those are the industries that have traditionally been a bit slower to adopt these types of technologies, but I actually think that's going to flip, and we're going to see a very rapid rate of change in those traditional industries as they adopt technologies.

Autonomous trucks, sensing, all those types of technologies, which will transform traditionally old industries and really improve their competitiveness rapidly.

What impact will AI have on jobs?

Lachlan: So you talked about trucks, autonomous trucks. 12 percent of the world's workforce is engaged in transport or logistics of some kind. How is artificial intelligence going to change employment. Is it going to destroy or create more jobs in the future?

Andrew: These are big numbers, and a million people driving vehicles sounds like a lot, and is a lot, but there are millions of jobs created every year and if you think about the time frame by which that gets implemented, it's a large number, but it's by no means an overwhelming number compared to the dynamism of the economy and the number of jobs that are created every year.

I think it depends a bit how you define a job. If you go, the work as it is today in 2017 is what we're going to call a job, then yes, there are going to be a lot fewer people doing those activities. If you're willing to have a bit more of an expansive view about what a job is and think about jobs that might exist in 30 years' time but don't exist today, and certainly weren't even contemplated 30 years ago, then I think humans are pretty good at developing new tasks for ourselves. Every wave of change has generated some displacement and ultimately, humans have found new ways to create jobs and employment.

Is this time different? I think it has different characteristics, but I think the fundamental ability of humans to find ways to deliver value to each other is not going away.

Lachlan: Elon Musk says robots and computers are going to be better than us at everything. What happens then?

Andrew: I just don't agree with that. I think there'll be lots of things that robots and computers are better than us at, but we have a comparative advantage in one fundamental thing, and that's being human. And I think we're always going to value that. And the expression of that through a job might change a lot. There'll probably be a lot less people doing routine tasks, a lot less people doing physical tasks, and a lot more people climbing up Maslow's pyramid, delivering the types of tasks and services that make us human, and that's something that we're well-equipped to do for a long period of time.

Niki: There's also a false choice between human and machine, and I think throughout history, it's the human enhanced by the machine that has always won that equation. So I think it pits it against this Hollywood movie of machine versus human, when in reality, people are made smarter and amplify their output through the use of a machine or a computer. And so I think if history is any guide, that's just going to happen more so in the future with AI.

The other thing is, everyone used to work on a farm. Then everyone used to work in a factory. And again, it's gut-wrenching in the short term of someone losing their job or having to retrain and find another area to use their time with, but over a longer period, I think it's just so inevitable that people find more and more useful tasks with their time when they don't have to do the base level tasks.

Which jobs are most at risk?

Lachlan: So the speed of the transition obviously matters a lot here. If it happens very quickly, then we're more at risk. Which jobs do you think are most at risk in the shorter term, Andrew?

Andrew: In any job that is composed of tasks that are able to be automated. And the tasks that are more able to be automated are the routine tasks and manual tasks. And so there are bunch of jobs that are clearly at risk because they contain a lot of those tasks.

However, the important thing to say here is there is no evidence right now of a big acceleration in the rate at which automation and artificial intelligence is killing jobs. This year, about 0.5 percent of all jobs we lost to automation and productivity enhancing machines more broadly, that is no higher than it was in the 60s, when tens of thousands of jobs were being lost in agriculture from the tractors and other farm machinery displacing labourers in agriculture. It's no higher than it was in the 80s and 90s when machines were displacing labour in factories around the country. And you only have to look at the productivity statistics to see why that's true.

We do not have ... Globally and in Australia, relatively low productivity growth, and that is not evident of a huge wave of machines displacing humans.

So if we're headed for some hockey stick change where this accelerates really rapidly, that's possible, but there's definitely no evidence of it right now.

Which companies are early investors backing?

Lachlan: There's a lot of discussion about which industries will be disrupted and which companies, but maybe flip it around for you, Niki. You invest in companies, you've got to look for which ones are the easiest for you to disrupt. Where are you putting your money right now?

Niki: I think more broadly, it's robotics. I think the most interesting example of that is autonomous vehicles, so the taxi or the ... It looks like the taxi industry but is actually the people owning a car or just pressing a button paying for a ride in a car. I think that will have the biggest near-term impact.

Any kind of robot that needs to interact with unsafe areas in factories, automating a warehouse, automating particularly logistics and so on. All of these areas are ripe for this technology to be so cheap and so affordable that it's ready to be adopted in the next 12 to 24 months.

Winner and Losers on the ASX

Lachlan: So now we might just go through some winners and losers in the Australian share market. Just a caveat first, obviously this is a long time coming, so this is going to take a long time to play out, so this is probably not a call in the short term for any of these companies, but just interested in your perspectives on whether they'll be beneficiaries or losers from artificial intelligence generally.

Telstra (ASX:TLS)

Let's start with Telstra. Niki?

Niki: Telstra. I think Telstra, again, has this wonderful opportunity ahead of it. However, if it doesn't embrace something by the fold, or it does nothing, it will be deeply negatively affected. I think in the short term, I wouldn't be too worried. I think businesses that hug regulation and are more tightly regulated tend to survive and not be affected by technology change as early as those unregulated industries. So has an opportunity, hopefully ... I'd be optimistic, I'm a forever optimist that they would take some of that opportunity, but obviously, if they do nothing, then it won't end well.

Lachlan: Andrew? Quickly, winner or loser for Telstra?

Andrew: I think lots of upside if they're there to take it.

BHP Billiton (ASX:BHP)

Andrew: Clear winner I think. Lots of scope to really transform the mining industry. BHP doing a good job, close to the forefront of the technological changes in that industry and pretty ambitious plans to go much further.

Lachlan: Winner or loser? BHP?

Niki: I think winner. Obviously the opportunity to make their operations so much more efficient and the ownership of the rights and the scale of their production and the barriers to entry, they're in a nice position.

Healthscope (ASX:HSO)

Lachlan: And you talked a lot earlier about deflation in health care. How about Healthscope?

Niki: I wouldn't be able to comment specifically, 'cause I don't know too much about Healthscope, but I would say all health care companies are in danger of great change. I think you can't rely on the government debugging the bad incentives in the health care system. You almost have to make it free so that it eliminates all of that and so this bottom-up, free alternative to an existing system. Not tweaking an existing system, but creating a bottom-up, almost free system that can exist, and I think all levels of health care are in danger of that.

Lachlan: Healthscope, winner or loser?

Andrew: I think it's a very competitive sector and I think a lot of water to go under the bridge in this sector as to how this plays out. But yeah, again, lots of opportunity and up to them to seize it.

Woolworths (ASX:WOW)

Lachlan: Okay. And another business that you know a fair bit about, given where you used to work at Wesfarmers. Woolworths?

Andrew: Look, I think that retailers have a huge amount to gain. Targeted marketing, the use of their loyalty programmes. There is a massive amount of data to completely change the shopping experience and the interactions with the consumer. The question is, can they get up the curve fast enough and do they have enough data to compete with some of the incoming players, including Amazon but not limited to Amazon. And there's a real opportunity for the Australian firms, I think, to have a bit of a battle with Amazon using their much larger Australian data set. And that could be a real source of competitive advantage.

But remember, on the flip side, Amazon has got a pretty awesome competitive advantage in the data that they have on the product side. So I think we're about to see a little war play out between two different sources of data, duelling against each other.

Commonwealth Bank (ASX:CBA)

Niki: CBA, I think its fate will be entirely tied to everything that's not technology. It's obviously an all-in bet on the Australian housing market, so I think how that plays out… it may go positively, it may go negatively, but I think technology is such a small impact on it, its fate is already locked in to whether Australian housing does well or badly.

Lachlan: Interesting.

Andrew: I think CBA are certainly using data really well, and right at the forefront of the banking industry, and trying to use that data to support their customers and deliver services to their customers above credit, and I think that's got the potential to be a real competitive advantage for them.

Lachlan: Excellent. And so of the nine percent that you mentioned that are using automation right, do these five companies sit in the good nine percent or the other 91 percent?

Andrew: Overall, Australia under-indexes in most industries. The industry where we do best in this field relative to others is in mining and energy. And that's where I think Australia is at the forefront of the application of these technologies within industries, and that's a good thing cause they're areas of real strength for Australia and always have been.

Lachlan: Excellent. Niki, Andrew, thank you very much for joining us today.

Not an existing Livewire subscriber?

Register on our website to access all Livewire content and to start receiving the Trending on Livewire daily email