The following is a lightly edited transcript of a speech delivered by the CEO of CSIRO’s Data61, Adrian Turner, on July 31st at the 2018 Australian Financial Review Innovation Summit.

Adrian Turner will be delivering a keynote address at D61+ Live. Hosted by CSIRO’s Data61, D61+ LIVE will attract around 2,000 delegates from all areas of the Australian research and development ecosystem, including government, corporates, high growth SMEs and academia. It’s free. Register here. D61live.com.au

A banner showing D61+ Live information

This issue is Australia’s number one priority. There was a lot of talk yesterday about AI and the potential for AI in the country, but if we don’t get this right then it doesn’t matter – we won’t be able to realise that opportunity.

The Model T Ford (Wikimedia Commons)

Let’s go back before we move forward and think about the car. The reason I bring up the car is because it is a general purpose technology. This means it’s a technology that impacts every industry. The first production car was developed by Karl Benz in 1885. Ford developed the Model T for the first time in 1913 and by 1927 they had produced fifteen million of them. What’s interesting about the car is not just the technology behind it, but the impact that it had economically and societally. There were many derivative industries that were created around the car. A lot of infrastructure that needed to be built – there was urbanisation as people moved to cities. Laws and legislation emerged around safety, and ultimately financial services products emerged as well.

A more recent example is the smartphone. Innovation and breakthrough technologies are often combinatorial. They build on prior cycles. Apple didn’t invent the touch screen or really any of the components, but they put it together in an innovative way and that led to the company going from near death and pundits writing the company off to shooting towards a trillion USD market capitalisation.

Why is there so much interest in machine learning and AI? It’s another general purpose technology that has the potential to generate a lot of impact, as the car did in the past. It’s combinatorial because what we’ve got is structured data. We’ve got Amazon Web Services and near infinite computing power on tap. And we’ve got algorithms developed in the 1970s – neural network algorithms – that for the first time have the data required to train them and the computer infrastructure to really crunch through the algorithms.

This is all coming together, and it will have a profound implication. What we’re seeing is new economics emerge through the proliferation of data, computer infrastructure, algorithms, and there’s been a lot of talk about platform players. We have 7 of the top ten by market cap platform companies. The top 5 are worth $3.4 trillion market capitalisation.

This matters because the tech companies that have got a jump on Australia and other parts of the world and are ploughing the money back into R&D. Tim Reed stood up yesterday and told us that  MYOB is increasing R&D investment from 15 to 20 percent. We’ve got the top 5 companies alone investing US $75bn in R&D.

Digitally, we see around the world that 11 per cent of GDP is now attributable to digital technologies at a whole-of-economy level. This has increased 10x since 1980 in terms of the proportion of GDP. But we think this is underrepresented. It’s a different story if we bring that global context to where we’ve got combinatorial innovation, new innovation building on prior cycles. We should take into account that Australia has substantially missed out on the first two tech cycles from an infrastructure point of view and from a platform company development point of view. As an economy, we’ve benefited in the past from productivity gains from the Hawke-Keating reforms. We’ve benefited from terms of trade but we’ve got a hole to fill and the only way to fill that is to create new industries and help existing industries become global. These will be underpinned by technology and data.

If we break it down into the contributors to digital value in an economy, there’s four types. Productivity gains (think about automating parts using software – relative to our OECD peers we’re at about 75 percent) are not yet at parity with our advanced OECD peers but we do reasonably well in applying tech to productivity improvements.

The second area is investment in digital capital. This includes things like know-how. Do we understand how to value data, how to wrangle and secure it properly?

The third area is creating domestic industries using data and digital tech. Most important for the future success of Australia is being able to create export industries. Globally relevant, digital, scalable, franchises. And we’ve got some, but not enough of them. Right now we under-perform relative to our OECD peers in building these. This is the area we’re very focused on – seeding these new industries.

When we drill into it further, the things that drove Australia’s success in the past is based on a 1991 model used by a lot of countries around the world looking at drivers of competitiveness. Things like natural resources, cost of capital and proximity to customers are not the things that are going to be important in creating these new globally scalable export industries. Instead it’s things like R&D, skills, our management and leadership, understanding of digital and how to scale companies globally.

Bill talked yesterday about business investment in R&D falling backwards at exactly the time the country should be over investing in R&D to create these new industries. If you look at the digital and data component, we’re falling even further behind. It’s not just that the numbers are falling behind at a macro level or an absolute value. Our investment in R&D in digital technologies is falling even further behind from an industry point of view, too.

This is a big problem. We’re starting to see the ramifications of this, with businesses struggling. We’re going to see more international companies entering the Australian market and leveraging their scale. Think of Amazon, who use data from their global footprint to run concurrent experiments around the world for customer conversion. Amazon has 5 times the customer conversion rate compared to Australian online retailers. How do you compete with that?

About 20-25 per cent of Australian firms are ‘innovation active’ – they’re innovating on some level. Then you break it down and look at the type of innovation. 80 per cent are doing innovation that is new to the world, exactly the kind we need. And it shows up in our indices. There was a conversation before about the makeup of the ASX. We see 3 per cent of firms on the ASX being IT firms. If you look at our top five companies on the ASX, the youngest company is the Commonwealth Bank of Australia, started in 1911.

We think if we can close this gap and just get to parity with our advanced OECD peers across those 4 dimensions that I mentioned, that represents a $315 bn GDP opportunity for Australia over the next 10 years.

We think the opportunity for the country is around what we’re labelling digital + domain. It’s the application of digital and data technologies to the industries that we understand deeply. Digital agriculture, health, energy, minerals and exploration.

Through analysis, we’ve identified some of these areas for the country where we think we have competitive advantage in the industry sectors – we think we can accelerate the digital part of this and create new industries that will lead to economic growth and development. At the firm level, we allow firms to adjust and compete globally.

If you think about what’s happened today, it’s the industries that have been fundamentally information-based industries that have been impacted. It’s media, it’s communications. What’s going to happen over the next 15-20 years is that same dynamic is going to wash over every industry sector. We have an opportunity to get ahead of it.

Let’s look at some other examples. Data driven urban management. Cyber-physical security – securing remote assets. We’ve got world leading capability in securing cyber physical assets. Supply chain integrity and visibility. Our small businesses need to plug into value chains. Australian governments are good at opening and releasing data sets, ranking above many of our OECD peers. Legal informatics and exploration. These align nicely with our export markets today – 75 per cent alignment.

So if we bring that back – the macro view back to the firm, how does the firm deal with this? This is the dilemma right now.

If you’re an incumbent company and you’ve identified an opportunity, it’s incredibly difficult to get there because of existing organisational structures, access to capital, talent, market, skills. If you’re a new company (NewCo), it’s getting easier to access the enabling technology to go after that opportunity. In many cases the NewCos are getting there faster. It means that structurally, incumbent companies have to first recognise the opportunity, set up the internal structure to allow a team to form and drive the change through the whole organisation.

There are two gains going on in parallel. The ones investing in digital technologies and embracing it that are going to win. At the core of this and underpinning all of this is data competitiveness. This is what we believe is the single most important thing for the country and industry to be focused on right now.

If we get really good and really competitive in the global context in the way that we deal with data, we can unlock value in these industries and also capture the value around ML and AI. We can also capture the value around the IoT. This includes things like data capture, data management and data analytics.

When we have a general purpose technology emerge in an economy, the benefit to the economy is proportional to the rate of diffusion of that general purpose technology.

While we need to capture productivity gains, we also tend to frame the opportunity based on our past experiences. So we think of the car as a horseless carriage. Right now we think about AI and machine learning in the context of enterprise productivity software. Because that’s the frame. We’re not thinking about it in terms of new value creation. What are the things we can do now that weren’t possible? Not just substitutive automation, but complimentary automation?

Look forward 10 years and ask what your industry will look like 10 years from now. What are the core enabling technologies that we should be investing in through R&D?

There’s a $1.6 trillion global R&D spend and we think there’s an opportunity here for Australia to get more than its fair share of that. If you think about what Facebook was or is, it’s fundamentally a new type of communications platform. There’s a technological shift going on right now, moving past the centrally vertically integrated platforms to more decentralised federated models. This is an opportunity for Australia.

This includes things like supply chain integrity around food provenance. Around 50 per cent of beef products are switched out for counterfeit products on the ground in some other countries in the world. That impacts our brand – brand Australia – and it impacts our ability to command a premium for our exports. We can’t outspend.

We’re very good at research – we’ve got smart people. Where we’ve fallen down is the ability to commercialise and scale that up. We’re not going to outspend other countries.

Our dilemma is that we don’t think big enough and in national terms from a research perspective.

Take robotics for example. We have so many research groups but none of it is nationally coherent. As a result we don’t do ourselves justice in presenting usable, scalable research that can be applied to industry. This is also an opportunity for the research side, research organisations in the country to get more linked up. This is where Data61 has been focusing its time, in building a new R&D model, applying platform thinking into a network.

We work with corporates to be an extension to their R&D organisations, we work with academia too. We have a single collaboration agreement across 30 universities, as well as government and investors. This model is now being replicated by 3 other countries, based on Data61.

I want to leave you with one thought. We started talking about combinatorial innovation. Just to contrast how some others in the world are thinking about this – let’s take an autonomous vehicle. Think about what goes into it – sensors, computer vision, communication systems, cyber physical cybersecurity.

By coming out and saying we’re going to be a leader in the world for autonomous systems, a country gets very good at all the subsystems that then can be applied to every other sector of the economy. Robotics, health, agriculture.

This is starting at the other end, thinking about combinatorial innovation and how we get good at the sub components, then get good at other areas. Guess what? This is what China’s doing. They want to be the number one in the world in autonomous vehicles. This is why you see all the partnerships with Western firms, all the investment.

What’s our moonshot? Precision health, agriculture, robotics. As a country we need to rally around industries we can be the best at.

Read more about D61+ Live, check out the agenda and register your interest

Adrian's speech will be at 11:30 on day one of D61+ Live, on the 18th of September 2018. With keynote presentations, panel discussions and 50+ projects on display, D61+ LIVE is an opportunity to meet the nation’s best data scientists, engineers and experts on all things data-driven innovation and discuss ideas and solutions face to face.