The world is currently amid an AI revolution, with rapid advancements in the capabilities and applications of AI opening the door to new innovations and opportunities. Despite being coined over six decades ago, the term ‘AI’ has undergone drastic acceleration via the elaboration of machine learning (ML) in recent years, said Puneet Mishra, senior scientist of AI & data modelling at Wageningen University & Research, speaking at F&A Next last month.
Unlocking global optimisation of food supply chains via AI
As AI continues to advance, it presents an opportunity to optimise processes, drive efficiency, and tackle challenges faced by the food industry.
Giuseppe Lacerenza, principal of Slimmer AI, a Dutch venture capitalist fund investing in B2B SaaS AI companies, said: “The real opportunity I see is moving from what has previously been local optimisation, to what we can call a global level of optimisation, from farm to fork. The opportunity sits in increasing the scope of optimisation and being able to optimise where humans haven’t been able to do so before.”.
This shift from local to global optimisation may allow for unprecedented advancements and efficiency improvements across the supply chain.
AI’s role in building the ‘farm of the future’
One key area where AI is making strides is in the development of the ‘farm of the future’ and autonomous greenhouses, said Mishra. By training robots to operate within farms and greenhouses, AI automates decision-making processes such as harvesting, and the sorting of crops based on quality.
According to Mishra, autonomous greenhouses and cultivation could be a significant area of AI-fuelled innovation causing a paradigm shift in the next five to 10 years.
By leveraging AI, the industry can “optimise the cultivation of whole crops in indoor environments to boost productivity and ensure the highest quality [yields]”, he said.
Boosting efficiency of agrifood businesses
AI's potential to enhance efficiency in the agrifood sector extends beyond farming and cultivation. The use of sensors combined with AI algorithms for example, could also optimise manufacturing processes, Mishra explained. By targeting quality and consistency, AI could help to streamline operations and optimise productivity in the manufacturing phase.
AI may also play a part in boosting efficiency within food businesses, drawing on capabilities such as enhanced search experience. Applications like Chat GPT, which according to analysis by Swiss bank UBS is now the fastest-growing app of all time, can increase the efficiency of tasks, leaving space for more added value tasks and creativity.
“Organisations will need to rethink the way they’re organised to leave more time for creativity,” said Lacerenza.
Adobe / mdisk![ai-opportunities-food-businesses.jpg]( "ai-opportunities-food-businesses.jpg")
Data availability is an obstacle to AI adoption
The availability of structured datasets has consistently been a significant barrier to AI adoption in the agrifood space.
“AI is very active in certain sectors such as the finance and medical sectors because there’s [an abundance of] structured data. However, [the lack of] structured data in the food and agriculture sectors, has limited the acceleration of AI in this field,” Mishra said.
Lacerenza agreed: “The availability of data has always been so far the biggest bottleneck to adoption of AI,” he said, explaining that the reason for this lack of data is often data storage issues.
Data storage issues, combined with legacy systems in large companies, pose significant challenge to the adoption of AI within the food sector. Startups, on the other hand, have the advantage of setting up their own data architecture, making it crucial for established companies to adapt and standardise their data management processes.
“[Data availability] is a need and a must to the value that AI can unlock today,” Lacerenza said.
In future, AI itself may provide a solution to this problem by generating missing datasets and enhancing the opportunities for innovation and automation, Mishra pointed out.
AI requires regulation and transparency
Regulation is both a barrier and a necessity to the future development of AI, the experts agreed. According to Mishra, regulation is essential to ensure that the outputs generated by AI systems and reliable and meaningful.
“AI cannot just generate anything; it needs to generate things that make sense,” he said.
Ensuring that training processes remain transparent is also critical to avoid biases and ensure control over AI technologies, as well as to mitigate risks.
“Regulation needs to really push towards the explain-ability of AI as that will give us control over a super helpful tool that today can come across as a bit out of control,” Lacerenza said.
