Jonathan Richetti1, Roger Lawes1
1CSIRO
A large language model like ChatGPT did not write this, or did it? Recently, Artificial Intelligence (AI) received another boost in interest and speculation. However, what it actually can do for agriculture is still mostly obscure. The goal is to provide a better understanding of basic concepts of AI using simple language and a practical example. We explain how to use AI in tabular data and provide guidelines for new users in the agriculture community. By doing so, we hope to narrow the knowledge gap between AI specialists and the agriculture community, including agricultural scientists, agronomists, and students. We present a historical contextualisation and de-jargonise artificial intelligence in relation to agriculture. Clarity for new users around the use of AI techniques to solve agronomic problems is provided. Key AI concepts are introduced; best practices for data pre-processing steps and metrics are recommended. Cross-validation is clarified, and its importance in agriculture is highlighted. It is shown that AI performance can vary with architecture and that the optimal choice is task-dependent. Emphasis on practical aspects for applying AI models for agricultural datasets is provided. Furthermore, examples of the use of AI in agriculture are further explored, such as detecting paddock boundaries from satellite imagery, decision-making processes in precision agriculture, parametrising process-based crop models, understanding crop rotation impact on yields, and determining grain flour quality. We provide guidelines for using AI techniques with a case study using different models/methods to forecast yields in cereals. Lastly, we discussed various issues and potential pitfalls of the misuse of AI in agriculture alongside potential benefits and their implications for the sector.