Unlocking the potential of AI innovations to help reshape New Zealand farming

Large Language Models (LLMs) like ChatGPT, Electra, and T5 are at the forefront of this transformation, and promise profound benefits for New Zealand farmers. However, the full extent of the benefits of AI remains, as yet, largely untapped.

A research collaboration between the University of Canterbury’s School of Product Design and AgResearch is exploring whether Chatbots, which are powered by LLMs, can accelerate change and whether farmers are also ready to place their trust in AI tools. 

The work is being led by PhD student Mamehgol (Mimi) Yousefidashliboroun under the supervision of AgResearch Senior Science Engineer, Mos Sharifi and Senior Scientist, Alvaro Romera. 

Mimi embarked on her project by engaging with dairy farmers across Southland, Canterbury, and Gore regions. Her aim was to gain insights into how farmers navigate the vast troves of data generated by modern farms, spanning weather forecasts, soil moisture levels, irrigation needs, livestock health, and overall farm management. 

Mimi says: “A consistent theme across the interviews is the balance between the wealth of experience and intuition farmers have and what technology's role is. While farmers are using various applications as a direct input in their decision-making, more advanced systems such as intelligent decision support systems appear to be used by mainly farm consultants.” 

Farmers, and farm consultants, have a large number of digital tools at their disposal. Overseer, part owned by AgResearch, is one such example. 

During her interviews Mimi said some trends emerged in regard to barriers to the uptake of AI by farmers. 

She said farmers often develop a relationship and rapport with farm consultants. These can be crucial to developing trust and over coming barriers such as real or perceived time constraints and information overload, which hampers decision-making, and the uptake of digital technology adoption and operational productivity.  

“There’s a pretty complex interplay between data, technology, and expert knowledge in facilitating informed and timely actions on farms. That is why from the insights gained during the interview and workshop with users, we identified that there is a need for conversational agents.  

Enter conversational agents powered by LLMs—a promising solution that bridges the gap between complex data and actionable insights. These agents serve as recommendation tools for both domain experts and laypeople, offering interactive access to computational models and data processing solutions. 

The next phase of Mimi’s research will look at “how to create intelligent interfaces for conversational agents, designing interfaces that foster user trust and increase adoption in the NZ agricultural sector.”

As technology continues to evolve, the potential for AI to revolutionize farming in New Zealand is limitless. With researchers like Mimi leading the charge, the transition that agriculture will make will be smoother and hopefully easier. 
“That is the aim of the research. To develop and share an understanding with developers on how to create tools for farmers that will use and trust.”