Communicating the use of artificial intelligence in agricultural and environmental research

Abstract

Transformative technologies such as artificial intelligence (AI) make difficult tasks more accessible and convenient. Since 2018, the use of AI in research has increased drastically, with annual publication rates of 3–5 times higher than pre-2017. Currently, >100,000 manuscripts using AI are published annually within science and engineering, and >20,000 of these belong to the agricultural and environmental fields. Given the magnitude of use, clear communication on how AI is used and how it helps advance scientific knowledge is essential. Clear communication is perhaps more necessary with AI than previous technologies due to its broad and flexible spectrum of uses, the “black-box” nature of deep-learning algorithms, and ongoing debates regarding AI's predictive power versus knowledge of first-principles mechanistic and process-based theories and models. In this commentary, we provide guidelines and discussion points to the scientific community to ensure transparent and effective communication of AI research in agricultural and environmental research publications.