Artificial intelligence (AI) has boomed in the past decade as it has become more applicable to daily life and more accessible to the wider population.
In the world of agriculture, food and nutrition, AI is used in areas such as increasing agricultural yield, developing sustainable food sources and improving nutritional value, but what more can it do and how can companies and institutions use it effectively?
The different faces of AI
For Shanghai Synbio Innovation Center (SSIC) inaugurated just two years ago, AI is deeply embedded in every stage of its work.
Functional nutrition and health, alternative proteins and novel foods, animal nutrition and pet health, as well as sustainable agriculture and biological crop protection are its key areas of interest when it comes to food and agriculture R&D.
Across these areas, SSIC focuses on enabling technologies such as synthetic biology, precision fermentation, metabolic engineering and bioprocess development to accelerate innovation from proof of concept to commercial application.
An example of what the center is working on is single-cell protein produced through gas fermentation using carbon dioxide and hydrogen.
“The aim is to create an additional protein supply route with less dependence on agricultural crops. Our work is focused on validation and the pathway towards a scale-up,” said Becky Zhao, head of international cooperation at SSIC. “It illustrates the potential of microbiome protein for the food and feed value chain, with the eventual application depending on performance cost and the relevant approvals.”
AI is one of the main tools used to fuel the center’s work, spanning from identifying groundbreaking technologies to the brains behind it.
“At the center, we use AI at every stage of our work. We use AI to screen our technology to see how advanced this technology is and to identify which principal investigator or researcher from across the globe possesses that technology,” she said.
This is only the first round.
AI is also used to screen and map the information gathered from the first round of screening into the center’s internal system – which currently has a database of over 580 technologies – and the team behind these technologies. The goal is again to identify and prioritize technologies that the center can support.
While AI could speed up the process, human judgment should not be omitted.
With the data gathered, the center taps into both internal and external experts – including representatives from the industry and regulatory institutions – to assess the commercial value, potential and feasibility of scaling the technologies identified.
“An AI platform can optimize the process and shorten the validation time. It is like a cycle – we build, we test and we learn. AI can enable the building of the cycle and to speed up that cycle,” she said.
So far, the center has also supported the development of 14 technologies, including AI-enabled technologies such as digital twins.
“We have already supported a digital twin technology which is designed to combine protein language models with models of human metabolism,” she said. “It is developed to start with population-level assessment of how foods and drugs are metabolized, with a longer-term ambition to predict individual responses.”
Virtual cell and organoids are other examples of AI-enabled models that the center is interested in.
AI for democratizing nutrition
There is a lot to be said for the role AI and other forms of technology play all along the agrifood value chain from production to manufacturing to innovation, but at what point does this translate into benefitting consumers from a health and nutrition standpoint?
According to CJ CheilJedang Executive Advisor and former SVP and Managing Director of Oceania Eugene Cha-Navarro, the litmus test for any technology and whether it can truly make a difference in this space lies first in its ability to shorten the product distance between the lab and the shelf.
“I come at this from the commercial side [and] where I see technology earning its place is in the unglamorous middle of the chain,” she said.
“[We know that] reformulation tools that let a company take sodium or sugar out of a staple without losing taste [and] demand sensing can tell which pack sizes and price points actually move in a traditional trade outlet in Ho Chi Minh City versus a supermarket in Melbourne. But the bigger point here is that the technology for better nutrition largely exists — But what is missing is access to it.”
This adds to an ongoing narrative that many healthier products are available in the market, but many of these have been designed to target a premium demographic, leaving market access to the average consumer from either an affordability or accessibility perspective lacking, to say the least.
“Here I think the food industry can learn from Korean cosmetics. K-beauty did not conquer the world because every brand owned a laboratory. It did it because a handful of ODM manufacturers made world-class formulation and production capability available to hundreds of brands, so a small company with a sharp consumer insight could get a proven product to market in months,” Cha-Navarro added.
“Similarly, if we want nutrition innovation to reach the middle of the market quickly, that is the model to build: shared access to formulation, efficacy testing and flexible manufacturing, so that the brand can concentrate on the consumer.”
Many companies fear relaxing their grasp on formulation and product development technologies for fear of losing out on profit margins, especially if products are eventually lower in price to access a wider consumer base — but if the growth of K-beauty is anything to go by, there are benefits to be reaped here for the entire industry.
“The traffic also runs the other way. Ingredient science developed for food is finding its way into adjacent categories. In Korea, probiotic strains built for dairy and functional drinks are now used in skincare, and the line between what nourishes from the inside and what nourishes from the outside is becoming a commercial opportunity rather than a category boundary,” she added.
“A manufacturer that understands its ingredients deeply can extend their value well beyond the original product.”
Bringing in the concept of AI as a driver, Cha-Navarro is certain that this technology can act as a catalyst for growth, but it cannot replace human touch in a process, at least not yet.
“What I want to address on top is that data is most valuable when it is paired with judgement. Data tells you what people did. It does not tell you what they would do if the product were right. That judgement still comes from people who have stood in the aisle and watched a shopper pick something up and put it back,” she said.
“The organisations that get the most from AI are the ones that have built the habit of turning insight into a decision quickly. That is an organisational capability first and a technology question second.”
AI’s success: The role of deployment and execution
As AI becomes increasingly embedded across agriculture, investors are becoming more selective about the types of startups they back.
“We’re staying away from software that claims to be the one system to rule the world,” said Jasper van Halder, chief innovation officer at Ravensdown and chief executive of Agnition Ventures.
He pointed to AI chatbots as an example of products that claim to know everything from soil and weather to animal health, adding that AI technology has advanced so quickly that these capabilities can now be replicated by combining existing large language models and tools.
“A year ago, we were still all surprised by the magic they could bring, and now we think the development will go quickly, and it will become obsolete,” said van Halder.
While some companies point to proprietary datasets or access to specialised sources such as satellite data, van Halder questions whether these differentiators are meaningful enough to sustain a competitive edge.
“We’re a bit sceptical of anything that is related to unique intelligence or unique expertise. We think that will be absorbed by the aggregation of all data, all intelligence, and all research. Everything will go into one source in the future. If that’s your unique value proposition, then it’s probably not going to last very long,” he said.
Today, the ability to execute effectively is more valuable than having unique technology, said van Halder.
“For instance, in pasture-based systems, you might have wearables for cows, e-tags or similar technologies. The technology itself may not be that novel, but if you’re first to market and you have the deepest pockets, you can get to scale because then you have an advantage. It’s no longer the expertise, but actually the deployment and the strength of execution.”
According to van Halder, he has already seen farmers turning to AI tools such as ChatGPT to seek out answers, much like patients researching their symptoms online before consulting a doctor.
“I think the farmers are the least of our problems. They will use AI. They’re just a bit cautious in terms of whether they can trust it and whether it works.”
This development is set to change the dynamic between farmers and agricultural advisers, who have historically used their scientific credentials and specialist expertise to help farmers make decisions.
With farmers increasingly able to arrive at their own AI-generated conclusions about what their land needs, agricultural advisers are being pushed to provide deeper insights and more tangible value.
“You can totally see a world where the models know more than any human can ever know in the future, so therefore the farmers have all the knowledge at their fingertips. Then what becomes of the advisors in the middle? The conclusion for us was that the only thing that can help you add value is the context of that farm system. Context, relationship and trust. This is, for us, the winning formula in an era of AI,” said van Halder.
Looking ahead, van Halder predicts that AI will become pervasive in agriculture but will be largely invisible.
“At the moment, we still talk about different companies and features as an AI app or something. I think it will all happen in the background… AI can supercharge that in the background by giving the ideas and the decision rights to the farmer. But I don’t think you will see farmers walking around with 40 gadgets or screens. It is the technology that will actually drive decision-making on the farm and can ultimately also be deployed in machines to do it themselves.”
Zhao, Cha-Navarro and van Halder are set to take the stage at the APAC Agri-Food Innovation Summit in Singapore from October 27 to 29, where they will shine the spotlight on China’s biomanufacturing surge to nutrition security, corporate investment and start-up partnerships.
Expect to hear from Microsoft Asia, Google DeepMind, XAG, A*STAR and more as they dive into the topic of AI innovation for agrifood R&D during the 3rd day of the event.
More details on the agenda can be found here, and tickets can be booked here.
