Meet the Solutionists, with Mark Scott

Growing food is one of our most fundamental human activities, and it’s the core puzzle underlying our societies. As more people are freed up from farm work, our societies grow and we need yet more food. 

Today, that tricky equation also includes robots. Australia faces a labour shortage in the agriculture sector for a variety of reasons, and the already razor-thin margins on many farms are only growing sharper. Enter: AI, sensor drones, self-driving tractors and a whole host of other high-tech tools. 

But are they too costly for smaller farms? What happens when a multi-tonne robot dies in the field? And what kind of workers will these robotic farms of the future need to run smoothly?

Professor Salah Sukkarieh, the Professor of Robotics and Intelligent Systems, has been at the forefront of robotics for twenty years. His own machines have been used on Australian farms, and his sights are also set internationally, where he hopes powerful machines will help self-subsistence farmers keep up. He explains what’s happening on Aussie farms already, how farm owners are managing the tech transition, and shines some light on the advancements to come. 

Mark Scott  00:00

This podcast is recorded at the University of Sydney's Camperdown campus on the land of the Gadigal people of the Eora nation. They've been discovering and sharing knowledge here for tens of thousands of years. I pay my respects to elders past and present, and extend that respect to all Aboriginal and Torres Strait Islander people.

Mark Scott  00:31

Right now, somewhere in Australia, a solar powered robot is crawling through a field, deciding which plants live and which die. It's sampling soil, identifying weeds and running complex calculations that will determine what ends up on your dinner plate. While tools like ChatGPT are raising questions about the future of knowledge work, the same thing is happening to the agriculture industry. Robots are already reshaping the most fundamental human activity, growing food. So what does it mean for the human project when robots take up tools on our farms? And what will it mean for you when the food they've grown reaches your supermarket?

Mark Scott  01:27

I'm Mark Scott, and this is The Solutionists. Salah Sukkarieh is the Professor of Robotics and Intelligent Systems at the University of Sydney, and a leading thinker in agricultural automation. So, Salah, if you took me out to a farm today, what would I see robots doing already?

Salah Sukkarieh 01:49

Yeah, so I think when you think of agriculture, there are many different commodities, there’s wheat, there's horticulture, there's meat and livestock, so grazing and across those different commodities, probably robotics is, you know, on different levels. On wheat farms, we've got flat land, it's open sky, there's no one around for kilometres, what we're seeing is what was originally just tractors that were driven by somebody end up being auto steered. So somebody just had to sit in a cabin, and this thing would just go up and down. And now we're seeing autonomous tractors, so no one in the cabin, the ability to kind of just say go to point A, go up and down the rows, the other end of the farm, come back. So in those types of commodities where those lands are quite flat and easy to navigate, they're the kind of robots seeing. We're starting to see that now in horticulture as well, mainly in vegetable farms, kind of the same style of land, but smaller robots now, four wheeled robots, probably the size of a car, maybe a bit smaller, doing similar things, identifying different plants. And so we're still in the research areas there about what kind of robotics would work on those farms.

Mark Scott  02:58

So the autonomous vehicle is going up and down, ot's identifying what you're growing, what you don't want to be growing, the sample of the soils. What kind of information are you getting from the machine? And to what extent is that more insightful than a farmer would be able to do on their own?

Salah Sukkarieh  03:17

When you look at farming, there are many activities that are happening. So the initial formation of the bed, the planting of the seeds, and then as the plant’s growing, there's activities such as identifying whether certain areas need more fertiliser, more water, other types of nutrients. They've got to spray fungicide or herbicide. Those types of thought processes where they go out into the farm and they take a sample of different areas, but they can only do so much. They can only sample so much of the farm, and then make a collective decision about what they can do across the whole farm, all the way through to harvesting. And what we're seeing now with robotics is this introduction to certain parts of those operations. So for example, instead of just taking samples of four or five locations on a farm, you have a robot that's crawling up and down the rows and taking precision measurement information down at the centimetre level. It can do anything, such as measure what's going on in the soil. So macronutrients, water, pH, you can measure what's happening with the plant that you want to grow. Is it growing at the right rate? Do you need to do something to it? It's also identifying plants that you don't want, such as weeds, and identifying how many of them there are. We're getting to the point now where we're also identifying diseases or pests on plants. So it's taking that process of sampling information so that you can have better decision making processes and being at a higher resolution. So that's one part, and the other part is doing things such as the more laborious aspects, such as planting maybe or spraying or harvesting.

Mark Scott  04:49

It strikes me when you describe that first part, it reminds me of conversations I've had on this podcast about precision medicine that once we would prescribe a drug for someone, and it'd be very generic, but now the technology exists for precisely the right kind of intervention for that person with that disease at that time. Is this a pathway to precision farming?

Salah Sukkarieh  05:13

Yeah, I think it's probably not as complex in terms of looking at the biochemical kind of relationships and the specificity of a particular plant. But when you think about a million plants on a farm, and you can save 1% of chemical costs, or you can increase yield by half a percent, and if you can do that, so the way to do that is to do that per plant. If you want to try and do both of those, because you have constraints. You want to increase the yield, but you want to minimise the cost. And that's what the process is happening there through the Ag robotics - you're providing that capability of getting that precision information, but also the precision action as well. So you can say this particular plant will get 0.25 ml of whatever chemical, the one next to what will get 0.23, the one that... you can get to that level, if it makes sense from a cost perspective, yeah.

Mark Scott  06:03

And then on the labour impact, certainly, you know labour costs, even though plenty of people who work on farms will be working on farms of their own, I guess. But is there also a capacity to just keep going, that in a way, humans reach the end of the day, humans get tired. Does it make farming more of a 24/7 operation with some of this technology that's coming to bear?

Salah Sukkarieh  06:26

Yeah, I mean, there are thoughts, for example, what you're doing is extending the factory out into the open space. So now things become a lot more designed in a way that robotics can be easier to implement. And what you are seeing is the ability for these platforms to operate 24 hours, seven days a week, and do things and not get tired through the process. But if we take a step back, the reason why robotics are going onto farms is not because farmers want to go 24/7, it's because they're not finding the labour that will work on farms.

Mark Scott 7:00

Interesting.

Salah Sukkarieh 7:04

So it's a different model. And so either in horticulture, for example, they rely very much on casuals, and they don't get the casuals all the time. Or if you look at a grain farm or a wheat farm, these large farms, they're very much family owned and run. And again, labour is hard. So the bringing in of robotics was originally because labour was hard to come by. With the introduction of robotics, because of that, you started to get all these other benefits that were coming out of it.

Oscar Pierce 07:30

My name's Oscar Pierce, with my wife Susanna, we farm a small property east of Moree in northern New South Wales, and we grow a mix of summer and winter broad acre crops and run cattle. I'm going to guarantee to make myself very unpopular when I talk about what is the most important tool that we've adopted. And look, I think one of the things when, when we have modern discussions and debate about AI, and this concern that AI is going to take away jobs. Technology has been taking away jobs on farms for years, for decades. That's the only way that farmers have been able to compete, essentially keeping up with the declining terms of trade. We've had to replace costly or inefficient things with sometimes capital expensive, but cheaper to operate machinery. But the tool that I use every single day that is most important is good old fashioned GPT five. GPT has allowed us to replace more costly human provision of various services, and for me, it's actually agronomy.

Mark Scott  08:43

Oscar studied rural science at university, immersing himself in agronomy, the science of soil and crop management, but not enough to do it professionally himself.

Oscar Pierce  08:55

And so we used to pay about $30,000 a year for an agronomist who would come out and check the crops, give us all the recommendations for nutrition, ensuring that we had disease, pest management, all that sort of thing happening. Because I've got a base skill. GPT has allowed me to replace that commercial agronomy service. I now use GPT pre-loaded. We upload literally everything. So I've put every single document I can find about wheat nutrition, wheat health, wheat disease, and have put all of that into GPT and its ability now to take photos, to recognise and to help you, essentially, to augment my own agronomy skills. Means that we're using it for disease identification. We're using it to calculate thresholds for pests, and essentially it's replaced humans and a $30,000 a year cost is now instead my monthly GPT subscription.

Mark Scott  09:56

And so has it worked for Oscar?

Oscar Pierce 09:59

We've only been doing that for about eight months, but the simple fact is, last year was our best ever wheat yield, and we've got our pastures probably looking more clean than they have for years, and we're spending less on chemical. It's been a bit of a breakthrough in terms of helping with those management decisions.

Mark Scott  10:18

So you've worked on a specific example. You've created this digital farm hand. What does that do, and how does that help a farmer's life?

Salah Sukkarieh 10:28

The Digital farm hand was built as a piece of technology, so firstly, as a roving platform that was Solar Electric, that was small, lightweight, approximately 300 kilogrammes, that could operate for long periods of time. On a nice sunny day, it would operate for 24 hours. It would get enough solar power to be able to regenerate its batteries and move on continuously. Once you have a platform that can move around like that, and you can operate it autonomously, then you can start to add tools and sensors onto it to do certain things.

Mark Scott  10:59

And to what extent is there now you know a commercial market that's delivering these kind of products and making them available at an affordable price for farmers?

Salah Sukkarieh 11:10

What we're seeing is the business model has flipped. So now, instead of a farmer worrying about buying or purchasing an asset such as a tractor, they would end up almost like leasing it, or even going to a robotics as a service type of model. You need weeding on your farm, and so calling in the technology to come along and do the weeding, just like you would call in casuals or contractors, is one of the process models that we're seeing, another one is leasing it. And so what happens then is the industry is taking on a lot of the initial absorption of the cost of building the hardware and the software and the expensive units, and then leasing them out or applying them as a service. So the models are changing. That's how it's become more affordable for the farmers. Farmers will only buy a tractor once a decade, once every 15 years. So being able to bring in, treating the robotic workforce, I guess, as casuals, is probably how we're seeing a lot of these models, business models turning.

Mark Scott  12:10

And if you're a farmer who's worked on the land for a long period of time, what new skills do you need to develop, or do you need to be able to access in order to run a farm taking advantage of these kinds of technological opportunities?

Salah Sukkarieh 12:25

If you look at some of the large scale farming operations, they have the capability to bring in house mechanics and electricians and software engineers that can support the introduction and the use of the technology on their farm to make it work within their operation. So introducing a robot is one thing. Where do you place it? When does it come out? How does it fit amongst all the other assets that I'm doing, whether I'm irrigating or sending out another tractor or whatever it might be, that requires some in house activity. And so in those particular areas, we're seeing that type of process happening on the middle size farms. What we're seeing is this robotics as a service model more so. So they're not they're less interested in the technology coming onto farm. They're more interested in the technology doing something on their farm. So calling in the contractors to come in, very much what we've been seeing over the last 10 years where farmers would call in drone operators to fly drones just collect data of their farm. It's a very similar type of activity. And then for the other 80% of the farmers that are still looking at this technology, it's really going to come down to when does the cost of the technology come down to such a price that they can use it.

Oscar Pierce  13:39

Nearly every day, you'll find some new application or tool or something that you didn't know. So that is, you know, that's great for the intellectually curious. I think the big thing is allowing farmers to do more with less, which has sort of been the mantra as farmers have to adapt to these declining terms of trade. Agriculture is no different to any other business. We're learning as we're doing somewhat as I’m testing capabilities and as new capabilities arrive, I am sharing it with other farmers, other you know, individual farmers, agronomists, we're looking at, you know, we have discussions on Twitter about what it can and can't do and so together, as a bit of a farm community, we learn more and more about where it's good, where it's bad, how to use it.

Mark Scott  14:30

And one of the key learnings Oscar and his colleagues share is that non-human workers simply don't have human problems.

Oscar Pierce  14:39

Those sorts of systems allow a spray rig to go out using integrated weather monitoring, so they always get the perfect timing, the perfect weather conditions. So those sort of technologies where a robot can do it better than a human, getting the environmental and operational conditions just perfect, and they can do it seven days a week, 365 days a year, because they don't have all the human problems. They don't get sick, they don't have kids soccer games to go to the weekend or other family life or social commitments. They're happy to come to work for 35 minutes and do the job, and then the weather window closes, whereas a normal staff member will not drive an hour into work, fire the tractor off and then stop instantly because the weather's not correct.

Mark Scott  15:33

This is cliche that farmers are conservative, kind of pretty set in their ways. This is a revolution that's coming to bear great opportunities. Who do you think is taking advantage of it and leaning into it? And what do you deal with those who just seem overwhelmed by the challenges and the opportunities of this kind of transition?

Salah Sukkarieh  15:56

I guess conservative is one way to look at it. Pragmatic, practical is probably another way to look at it. Ag has been around since the dawn of time, and so farmers have always been looking at technology of some form to help them do whatever they need to do on farm. And so if you bring a robot onto a farm, they're going to kick its tyres, they're going to push it around, they're going to see whether it's robust enough. And that's probably, your initial impression might be, are they sceptical about the technology? But really, at the end of the day, it's a it's a business that they're running. It's probably it's been in their family for years, and somebody coming along and saying, Here's a new piece of tech, they're going to really look at it carefully. So that's the mindset that I've seen across all types of farmers. Are the ones that are early innovators, the ones who kind of take it on board very quickly, usually other large scale farmers, ones who are very operational savvy, who have seen how technology has changed the way things have happened on farm, whether that's communication or whether that's in the shed and robotics there, or whatever it might be. But that doesn't mean that they’re also naive to the possibility of it not working, or anything like that, they still scrutinise and test it out. But they also become your vocal point, once it works, or once they see the potential of research and what it can do, they become the mechanisms by which the rest of the farmers follow up. So I think across the board, what I found, I mean, the farmers are systems thinkers. They get to encapsulate a lot of information. And you may think it's just a gut call, but there's a lot of experience there. And so whenever a new asset comes in there, a new robot comes in, they look at it carefully, and they're willing to take it on if it works.

Mark Scott  17:31

Agriculture is such an important sector in Australia, but in many countries that have had a big increase in per capita GDP, the relative importance of agriculture to that economy has diluted as other parts of the economy have grown. But in poorer parts of the world, agriculture is a vital industry to feed the country and also an opportunity for export. What does this technology mean in poorer communities than you'd see in Australia?

Salah Sukkarieh  18:04

Yeah, this is an area that, for me, has been the area that I've wanted to try and focus on a lot more of the last decade, which is, if you introduce this technology into the agriculture sector, then it's almost like successfully successful, right? You kind of, you're giving the technology to farmers that already know how to really operate and so a robot might come in and or AI system might come in, and then all of a sudden you increase the potential by one or 2% but the real big gains if you can go to a particular area and introduce this type of technology, and then you get a 30% increase in yield or less damage to the supply chain through diseases, pests, etc. And generally, what people feel is, you can't throw a robot into the middle of a third world country. The kind of process that you think through is, though, the farmer there in that third world country has exactly the same problems as a farmer in the first world country. They can't find labour, it's hard to grow food. Climate change is having an impact on pests and diseases easily coming through. They don't know about soil, so any form of technology can actually help them. The process is, how do you introduce that technology into that area? So we've had a number of trials, and we've gone to Fiji, Samoa, Indonesia. We've tested and recently, with some colleagues in agriculture and social sciences, we were awarded a grant with the Australian Centre for International agriculture research to test out a lot of these ideas in the Mekong region over the next five years. And when you look at it, the technology is the easy part. The hard part is the education, the training, the gender studies, the political analysis and how you wrap the technology within that more broader socio economic sphere. And that's probably a goal that we'd like to keep focusing on, because 80% of farmers are small holder farmers from around the world. That's that's where they are. And so if you can bring in that technology to help them, that's really where the greatest benefit would be.

Mark Scott  20:00

You've been to the cutting edge of robotics and intelligent systems for decades now. I'm curious what you think the next 20 years of farm technology will look like?

Salah Sukkarieh 20:09

3d printing and the ability to kind of build your own robot that meets your own needs, I think is going to be a big one. As solar energy gets a lot or solar cells become a lot more efficient, and batteries become smaller batteries, and the system itself becomes a lot more efficient over the next five years. We're going to see unique types of platforms. Soft robotics is going to be an interesting one. That's the ability to kind of build robotic platforms made out of materials that mimic a lot more of the flexibility of, say, a human arm, and what that means with harvesting, we're going to see a bit more of that happening. Quantum sensing is going to be another interesting area, because if its ability to kind of dig out information that we haven't been able to kind of pull out from both plants and soil, I think will be another change. So that's I guess on the technology front, I think what we're seeing also on, I guess, the operational commercial front is there's going to be those large farms that just get bigger, because they've got the operational power to use the technology and make it work. And the question is whether we can also make the technology work for the smaller farmers as well, and how do we kind of play that game out over the next 10 years.

Mark Scott  21:27

That's Salah Sukkarieh, the Professor of Robotics and Intelligent Systems here at the University of Sydney, and someone who's sure to have a hand in how we play this game for the next 10 years. If you're curious now about the other end of the line of production, what you actually eat, listen to our episode with world leading nutritionists, Professor Stephen Simpson and David Raubenheimer.

Stephen Simpson 21:50

The problem is that processed foods are engineered by food engineers and corporations specifically to override those regulatory mechanisms that tell us when we've eaten enough, because eating enough is not a good business proposition for processed food manufacturing, eating more is a better proposition.

Mark Scott  22:10

And if you want to hear how the best minds in the world are tackling our biggest challenges, make sure you follow the show in your favourite podcast app, so you don't miss an episode. The Solutionists is a podcast from the University of Sydney produced by Deadset Studios.

The Solutionists is podcast from the University of Sydney, produced by Deadset Studios. Keep up to date with The Solutionists by following @sydney_uni Facebook and Instagram, and @sydney.edu.au on Bluesky.

This episode was produced by Liam Riordan with sound design by Jeremy Wilmot. Supervising producer is Sarah Dabro. Executive editors are Kellie Riordan, Jen Peterson-Ward, and Mark Scott. Strategist is Ann Chesterman.

This podcast was recorded on the land of the Gadigal people of the Eora nation. For thousands of years, across innumerable generations, knowledge has been taught, shared and exchanged here. We pay respect to elders past and present and extend that respect to all Aboriginal and Torres Strait Islander people.