Early detection of potential animal disease outbreaks, so crucial to containing them, has received a shot in the arm from artificial intelligence that can pick up signs of trouble way before a human could.
The implications for biosecurity, say the developers, are immense.
The program, Sentinel-AI, essentially finds patterns in biological data, and has huge potential to reduce the impact of animal diseases on farms such as foot-and-mouth disease (FMD) and Mycoplasma bovis.
The solution – believed to be a world first – has been developed by Otago-based vet science company Ingenum in partnership with Qrious, the artificial intelligence (AI) and data analytics specialist in the Spark business group.
Led by founder Dr Tom Brownlie, Ingenum has been building disease detection and prediction tools for the past five years.
Brownlie, a veterinary epidemiologist and data scientist living at Taieri Mouth in Otago, was working as a production animal veterinary surgeon in the United Kingdom during its 2001 FMD outbreak.
“I saw first-hand what it meant for rural communities and families,” Brownlie says.
“The sooner a disease like FMD is caught, the sooner farming sectors and government can take action to stop it in its tracks.”
New Zealand has so far avoided the FMD outbreak that’s causing havoc in Asian agricultural markets right now, but experts are warning against complacency.
Brownlie started Ingenum in 2016 after working in agritech and noticing the growing potential and effect of on-farm technology. He knew that data could have a massive impact on the future of disease prevention and on-farm biosecurity.
His range of experience – from being a production vet to lecturing in epidemiology – “allows Ingenum to straddle the technological, biological and traditional domains of agriculture”, he says.
Ingenum’s experiences with data and disease detection and its understanding of the agritech commercial value chain is complemented and enhanced by Sentinel-AI, the company’s powerful artificial intelligence and machine learning offering.
“The [Sentinel] concept is straightforward, but sophisticated in its execution,” Brownlie says.
The program detects anomalies in clinical sign data observed in livestock – think shivering, bloating, or depression.
Combined with data from government agencies, veterinary practices and agritech companies that routinely measure the quality of final products such as milk composition, Sentinel-AI produces insights about underlying patterns.
Comparing this data with normal seasonal data shows up anomalies. Mapping the results geographically helps those manning a disease response to make better decisions in real time.
“Ingenum is effectively bringing together government, veterinary and agricultural technology companies in a secure and anonymised way to make sure that we use every advantage available to us to speed up these diagnoses,” Brownlie says.
He notes the changes in agritech that have occurred on NZ farms in the past five years.
“On-farm tech is an embedded part of agriculture being embraced here now. We know in FMD, the models we have built show patterns of on-farm technology, so we built the Sentinel prototype in 2018.
“One of the most profound things we learnt is the wonderful mess out there in data, it’s a muddled structure,” Brownlie says.
That is where Qrious head of technology Chris Laing came in. He was brought on board for the AI expertise he could offer.
“We are working at the absolute leading edge of what is available,” Laing says.
“This will transform the global livestock industry and its security.
“We are working with Tom because he has the vision no one else has come up with.
“The technology has only matured in the past year and NZ is heavy in the agritech space with a lot of data coming off farms now, allowing capability in modern machine, training and learning – a set of software observing and bringing data all together.
“The outcome is an AI system that detects patterns too complicated and subtle for humans to detect.
“It is watching over NZ’s biosecurity 24/7, the equivalent of a human watchman.
“It will not always tell you what is wrong, but that something is wrong, showing a pattern of clinical science developing for human experts to investigate.
“This is a critical phase of Ingenum’s journey and a rare opportunity to move the needle on something as important as it is to NZ agriculture,” Laing says.
Brownlie says it is not about replacing the human custodians of animal health.
“We are giving them more support, so they are not making decisions in isolation.
“With my farmer’s hat on, I hear, ‘I’m not sure I’m comfortable with big brother watching over me and the health of my stock’ – but it’s not like that.”
“The signals at farm level are not subtle enough, these data are aggregate and certainly anonymised.
“It’s about using what we have got collectively to make a meaningful difference nationally to biosecurity.”
With its potential for a meaningful contribution to national biosecurity efforts, the technology is now undergoing trials in Australia and there are plans to take it global, moving on to the UK and Ireland.