Adrian Turner, CEO, CSIRO’s Data61, delivered the keynote address at D61+Live, which was held on the 18th and 19th of September, 2018, in Brisbane. Video and transcript below.

Transcript

 

What I’m going to talk through right now I think is probably the most important conversation that this country needs to have. And I think we’re at a fork in the road nationally and hang with me for the next twenty minutes or so. I’m going to walk you through some data and if I succeed in the next 25 minutes or so what I’m hoping to do is to convince you all that we need to start a collective movement to change the trajectory of Australia. So let me start with Data61 and our reason for being.

So Data61 solves Australia’s data-driven challenges that are underpinned by deep science and technology. We’re the data science arm of the CSIRO and what’s powerful about that is not only the 1,100 team members including affiliates in the network, 680-odd PhDs focused on data science, strong engineering product management, and user interaction design capability, but then we get the ability to link up with and tap into another five and a half thousand talented team members inside of the CSIRO with enormous amount of experience across domains like agriculture, health, minerals, energy and access to about 2,800 industry partners. It’s a powerful combination. So what’s going on right now and Kate alluded to it, is we’re seeing the power of combinatorial innovation. So, in history we’ve seen periodically the emergence of general purpose technology. So a general purpose technology would be the transistor and in the transistor we see derivative products.

So pretty well everything with a CPU and a power supply is made possible today by the transistor. And then if you think about all of the derivative impacts and spillover on to different parts of the economy. So what you’re looking at here is Apple and Apple tracking towards being the world’s first trillion-dollar enterprise. And what really sparked this was the iPod, the network services that wrapped around the iPod that led to the development of the iPhone, that led to the development of an application development ecosystem, the iPad, so it goes on. But Apple didn’t actually invent any of the breakthroughs that form those initial devices. What they did was put them together in unique and compelling ways. And there’s an opportunity in this country to create the breakthroughs but also to think about other types of innovation and combinatorial innovation.

 

DARPA and Australia’s ‘Sputnik moment’

And why this matters now is I think Australia in the last weeks has had its Sputnik moment. So in 1957 the Soviets launched the first satellite and they beat the US to it and the US was surprised. And in response to that the US created DARPA. The Defense Advanced Research Projects Agency. DARPA is responsible for the internet, GPS and voice recognition initially. And what’s driving this change and I’ll come back to it at the end why I think it’s the last couple of weeks, is the emergence of a general purpose technology in machine learning and AI. And, AI is a lot of things.

So Kate touched on things like natural language processing as well as machine learning, computer vision, but this is a general purpose technology that’s going to wash over and impact every sector of our economy and the way that government operates as well. And it’s not just about productivity gains it’s about creating that new value. So what this is is and you see it, is Emersent, which is a spin-out that’s developed LIDAR technology combined it with robotics and autonomous systems technology to be able to self navigate in mine shafts and dangerous places where people shouldn’t go or don’t want to go. Incredibly compelling combinatorial innovation underpinned by breakthrough. Another example is we’ve got order of magnitude 5,000 sensors on the Harbour Bridge for structural monitoring for infrastructure.

What this is doing is extending the life of the bridge so what it does is in partnership with the New South Wales Government is its vibration sensing so when buses go over if there’s an outer band vibration it suggests that something needs repairing around the bridge. But it’s also environmental. So this is a project called Lion’s Share. This is a great example of using machine learning and AI not just for productivity gains but to create new value that wasn’t possible before. So this is an Australian group called Finch. They did, if you’ve seen on Netflix, Chef’s Table or Jiro Dreams of Sushi. They’re a production company. Academy award-winning production company. And what they came up with, or what the question that they asked was, ‘How do we give animals rights for appearing in TV commercials and films?’ It sounds like a really simple idea but in fact it’s only made possible by computer vision and machine learning to automatically recognise the animals in the film.

But what they want to do is to go the next step and say ‘well that’s not just an elephant that’s an African elephant’, collect the royalties and then turn that royalty stream back to protecting the habitats of the animals that are appearing. This fund, in the space of four months, and we’ve teamed up with them as their science and technology partner, has convinced David Attenborough to front it. It was announced at the Cannes Film Festival, there’s going to be an official launch at the United Nations in the next month, in New York, and this fund is going to generate a hundred million dollars inside of three years from these royalties. And those monies are going to be used to protect the habitats with things like we talked about, the Amazon biodiversity monitoring.

We’re also working with smart sensors to protect villages from elephants so they learn so they emit a sound and if the perimeter is breached every sensor now knows that that sound is no longer effective in that context. So it experiments with new sounds. Another example from the same group just thinking about voice is this group has come up with the idea of what about creating a voice bank for people with ALS that are going to lose their voice. So you think about you know the the very robotic you know Stephen Hawking’s sound for somebody who’s had ALS and loses the ability to talk. Well what if I could record my voice and have a computer system sound like me if something were to happen to my voice. It is powerful right and its net new value creation. So if we bring it back to the economics what we see if we move to data driven industry underpinned by machine learning. It’s a whole new set of economics kicking in and they favor scale global scale. So what we see now is seven of the top ten companies in the world being these data-driven platform companies. None of which have come out of Australia.

We see the annual R&D spend of the top five combined being seventy five billion dollars US and that includes product development from Amazon not just just R&D. To put that in context Australia spends twenty point seven billion a year on R&D. So just the top five tech companies are now spending close to three times what Australia does as a country and hoovering up a whole lot of talent to make that possible. So digital technologies are impacting every sector of every economy around the world and what it represents now is about 11% of GDP and growing. So this is what you see at the top in the coloured band is the GDP impact of digital technologies on economies around the world. That’s a six trillion dollar impact today globally.

 

Australia’s challenge

The problem for Australia or the opportunity is that we lag our OECD peers across four key dimensions of digital innovation that I’m going to talk about. If we only get to parity with those countries over the next ten years that’s a 315 billion dollar GDP opportunity for the country. Now the challenge we’re up against is that the drivers at a whole of economy level that led to the success and the 27 years of uninterrupted economic growth are not the drivers that are going to get us there going forward. So in the past it’s been things like natural resources cost of labour, cost of capital. In the future it’s things like how effective are we with our R&D and translating our R&D to create new breakthroughs that we take to the world that fuels economic growth domestically. What about our workforce skills? Do we have the right sort of digital literacy in our leaders and our teams? We think not based on some work that we’ve been doing with directors.

That there’s an opportunity to lift the literacy and a need to lift the literacy amongst our leaders. Also the management quality to be able to scale up so not just take an idea and turn it into a you know a Minimum Viable Product or prototype, but then turn that into a company and into global franchise.

So if we drill down these are the four dimensions. So the first one is productivity so this is applying digital technology to get productivity improvements. This is using technology and consuming technology. We’re not bad at that. We’re about 74% of our OECD peers. The next one is the investment in digital capital. So this is infrastructure, this is also know how, this is an appreciation of data as an asset. Here we’re about 71% of our OECD peers. The third leg is about or the third part is about domestically created digital industries and here we fall back a bit where it’s 67% of our OECD peers.

But the last one is the kicker that if we don’t solve this we will not enjoy the quality of life we have today. And we’re at 20 percent of our OECD peers in creating digital exports. Now we have to be a bit careful about averages you could argue that the problem may even be more acute than those numbers suggest. When we look at something like productivity and we take a whole of economy average, if we drill into the Delta between the companies that are embracing digital technology and innovation, they’re actually significantly more productive. 37 points more productive than those that aren’t, it’s just that we don’t have enough of those to move the national average yet.

And if you think about IT intensive service industries it’s plus 49 points of productivity improvement. So this shows up in our markets. So if you look at the ASX only three percent of Australian publicly traded companies are IT companies and if you look at our five largest companies the youngest of those is CBA that was created in nineteen eleven. So if you look at other economies around the world you get this healthy churning and cycling of new companies emerging. If you look at the five largest companies in the world you see Microsoft in seventy five, Apple in seventy six, through to Tencent. Amazon, Google as well.

 

AlphaBeta’s new report

All right, so that’s the current situation and the question is where do we go from here? What do we do about it? So what what this is speaking to is a report that we’ve released today that was commissioned by us, developed by AlphaBeta.

And what it’s doing is looking at some of these underlying factors, but then drawing conclusions about where Australia can go from here and where do we focus. And so what they did was they looked at these are the new technologies and what can be enabled. These are the domestic economic pressures unique to Australia, this is the social and environmental context for the country, so this is uniquely Australia and then overlaying that with potential places we could focus, what’s the significance, what’s the feasibility, and what’s the compatibility with our existing industry sectors and our economy. What they concluded was at the core of any of this transformation is data competitiveness, is national data competitiveness, so getting really good at how we capture data, how we manage data, including how we secure it, how we share it across organisational boundaries and jurisdictional boundaries, how we analyse and new techniques for machine learning and federated analytics, as well as how we make decisions so the human factors part of this. And what they concluded was there’s eight areas where Australia can really make a difference.

So, precision healthcare is one digital agriculture and there’s more detail in the report about the specific areas of health or Ag, data-driven urban management, so we do this today we do this well, cyber physical security so within Data61 there’s a trustworthy Systems Group, that’s actually teamed up with DARPA and done work with DARPA in the US to protect autonomous systems. Cyber-physical security is an opportunity for us building on our strengths in remote asset management in things like mining and other sectors. Supply chain integrity, so this is the tracking the visibility of goods and services through a supply chain, 98% of the Australian economy is small business having to plug into global supply chains. The only way we deal with things like counterfeiting of goods and services is by having visibility right through the supply chain, it requires new science and technology breakthrough. Proactive government, so the role of government as a business partner not a service provider. Legal informatics is taking law and legislation and making it machine readable so the compliance drag on the Australian economy is two hundred and forty nine billion dollars annually. Compliance software is one of the fastest growing segments in this country.

 

The economic benefits of innovation

In the US it’s a four trillion dollar drag on the economy. We’ve got unique capability in the country to develop new sorts of platforms and take those platforms to the world. And the last one is smart exploration and production for minerals. So using things like hyperspectral imaging to look for anomalies in in the land to help narrow the focus for more targeted exploration. So if we want to go after these or others, we have a challenge based on the scale of our economy that if these new industries are going to be underpinned by some of these general-purpose technologies we need to get very good at working with those technologies. So what you’re seeing here is just in the last 24 months other countries announcing national AI programs. So you’ve got the EU investing about 20 billion dollars, you’ve got France investing 1.8 billion dollars, you’ve got China investing about 30 billion dollars, the US is investing twenty to thirty billion dollars, Singapore is investing about a hundred and forty million dollars. And all are recognising that this is a game-changer for economies.

Now in the last weeks you’ve seen DARPA come out and challenge the whole notion of the scientific method in part by calling for submissions to build an AI in the next 18 months that can extract knowledge from scientific papers. So instead of starting with a hypothesis and testing that hypothesis the AI will bring back a hypothesis and say go further investigate this. It changes the dynamic. It has implications for how our research organisations are structured. Our ability to deal with data. So if we recognise that we need to get more organised to go after some of these emerging areas, how do we do it and how do we stack up? Structurally, how well-placed are we do this? So the first thing that AlphaBeta called out is our investment in R&D as a country is largely indirect. So about 13% of our R&D spend is directed. Where we say we want to be the best in the world at Quantum Computing. Or pick the area. By contrast the US is 75% directed. So the US says we’re going to be the best of the world in machine learning and AI. Israel is about 90 percent and recently in the last months you’ve seen Israel come out and say we’re going to be a leader in the world in precision health. It is a six trillion dollar global opportunity and we are going to capture 10% of that. Six hundred billion dollars, that’s the prize they’re going after. Germany is close to a hundred percent. So this system works where there’s indirect incentives, R&D, tax, R&D incentives if industry is leaning in and doing R&D. Except they’re not in Australia.

So the National, sorry, the OECD average is about 2.2 percent of GDP that businesses spend on R&D. In Israel its 3.6 percent, in Australia its 1.9 percent and it’s sliding. Our OECD peers are going up and to the right and we’re going down. So something is systemically broken. And we also underperform our peers in matching up research with commercialisation so we rank about 40th is that right, my eyes aren’t that good, about 40th. But it’s not all on industry we can’t say if this is a two-sided equation we can’t put it all on industry at all. So this is our research science and technology dilemma. So this is I asked the team I said pick an area it kind of doesn’t matter what one it is, could be cybersecurity, could be robotics map it. Map who around the country is doing what and this is what Australia’s robotics investment looks like today.

Lots of great smart activities going on but nothing of scale and no national coherence. And so what we did was we got nine of the groups around the table in the last three months and we all agreed that what was needed was more visibility across the research projects, we needed a national roadmap and we formed, Minister Andrews referred to, the Sixth Wave Alliance to project to the world the fantastic capability we have in this country around robotics. So if you put all that together and say well the industry side’s not firing on all cylinders, the research side isn’t firing on all cylinders, what does that mean? That means that only about 2% of our companies today develop new to the world innovation and then export it. And yet it’s the exporters that the fastest-growing in terms of job creation for the country and economic growth. So that number says eight percent but it it’s referring to innovation active of which only a subset of our companies are actually innovation active.

 

What’s Data61’s role?

Alright so if we bring it back to where we started Data61 – what’s Data61’s role in the system to solve Australia’s data-driven challenges underpinned by deep science and technology. I hope I’ve convinced you that we’ve got a massive one and data science is at the core of the solution. So for the last two years we’ve been quietly and methodically executing against the vision to develop the asset for the country that can make a dent and capitalise on this ‘delta’. So the first thing we’ve done is we said there needs to be a new model because the world and the context around us is changing so fast, and because we’re relatively small as a country, we need to apply platform thinking to research and redefine the way that research is done to be able to assemble teams around specific problems not just from within Australia certainly not just within CSIRO or Data61 but across the network of partnerships here and overseas.

That’s D61+, that’s what this event is about today to celebrate that. That’s a new model and we’ve got in alpha a capability directory sitting underneath that with 50,000 researcher profiles that with more investment and we’ve got about 40 data partners feeding data feeds or we’ll be feeding data feeds, there’s work to do on that but ultimately that will be an enabler to help remove friction from assembling teams. The second thing is our research focus, so what do we do? How do we apply that capability for the country?

And we focus, we’ve made a decision to focus around the four core areas that that that feed into or enable data competitiveness. So these map to the research programs that we have inside of Data61, so cyber-physical systems around collections, machine learning and analytics, software and computational systems is around data management and sharing, privacy, and the last one is decision sciences. The next part of the model is to say this vertically integrated platform model is relatively new, that’s led to this concentration of value and the creation of seven of the top ten companies in the world by market cap being platform companies. What we’re questioning is is there a different model does this evolve to be a different model where you have a more distributed federated value chain, where there’s more entities participating in that value chain. And we think that’s where this goes. So we think we’re heading towards a new class of entity that we’re calling industry utilities, shared infrastructure, data cooperatives, open standards-based with more participants in the value chain benefitting.

The Challenge Model

And we think applying the network, the focus, the research that’s enabling a new type of distributed platform, a new type of industry utility applied to those eight markets is how Australia closes that gap and realises not only the 315 billion dollar GDP opportunity but creates new digital exports. And to accelerate all of this we’re also announcing today modeled on DARPA and the Gates Foundation, a data challenge infrastructure and program for the country. And so what we want to do is move from individual you know subscale research projects to frame around what are some of the largest most important national scale data driven challenges and drop that challenge on top of this platform and at scale mobilise the entire system to respond to it. And the first one that we’re announcing and we’re going after with partners is food provenance and supply chain traceability.

So our Ag exports are incredibly important to the economy and it turns our counterfeiting is a big problem. So in some parts of the world as much as 50 percent of our beef products are switched out for counterfeit products. But also if you can trace through the supply chain you’ve got traceability. If there’s a food safety issue or a food recall issue you can trace right through the supply chain.

So our initial partners, KPMG, ANSTO, GS1, the AG and food group within CSIRO that’s globally recognised for their fantastic work, government, federal, state, and also Austrade, and there are others in there too, other industry partners that we’ll be announcing shortly. So this model is potentially a new model for the country to accelerate translation of research into outcomes. There’s 36 individual research projects that make up this initial phase of the challenge. If we get it right it also positions Australia well geographically. Because what’s going on in our region and in the world is these technologies, and I focused on the economic dimension here, but there’s also a geopolitical dimension as well.

Where the currency for our standing in the world will increasingly be our ability to create these digital exports and bring new breakthrough science and technology to the world and to our partners. So this is some work that we did looking at sunrise industries in the ASEAN, and just in the last day our CEO Larry Marshall announced the formation of CSIRO’s permanent office in Singapore. So we’re committed to the region we think we have an important role to play.

 

Door A, and Door B

Commercialisation pathways out, global context in, to make sure that we’re focused on the right problems and this is the last thing I want to leave you with, is, I don’t want to be alarmist about the urgency here but I’ve been back in country in this role now for three years, and I genuinely believe that we’re at a fork in the road where we walk through door a or door B. Door A says that we create a national movement and mobilise around creating new digital exports and everything that’s required to make that possible. Or, we walk through door B and for anyone who has kids in the room I think we have only one option which is to walk through door A. And it’s perishable because these investments are happening globally.

So in the last week aside from the science thing DARPA’s announced a two billion dollar program on next generation AI, talents leaving the country to go and be a part of programs like that. There’s knock-on effects for industry as well. So I don’t think this is a decision that can be taken in the next six months or 12 months.

This is a decision right now for us to mobilise around this. And we think we have a role to play hence the announcements today and we will continue to execute against that and we’ll continue to be vocal but we need your help and we need the help of others in the country to really succeed against this mission so thanks very much for listening to me I am going to be around and I’d be happy to pick up any part of the conversation over the next couple of days. Thank you.