Human versus machine: Can generative AI anticipate insect biological control outcomes?

Publication Type

Journal Article

Journal Name

Computers and Electronics in Agriculture

Name of Author

Kris A.G. Wyckhuys, Chrysalis Consulting
Komivi S. Akutse, International Centre of Insect Physiology and Ecology Nairobi
Divina M. Amalin, De La Salle University
Salah Eddin Araj, The University of Jordan
Marie Joy B. Beltran, University of the Philippines Los Banos
Ibtissem Ben Fekih, Université de Liège
Paul André Calatayud, International Centre of Insect Physiology and Ecology Nairobi
Lizette Cicero, Instituto Nacional de Investigaciones Forestales, Agricolas y Pecuarias
Marcellin C. Cokola, Université de Liège
Yelitza C. Colmenarez, Universidade Estadual Paulista "Júlio de Mesquita Filho"
Kenza Dessauvages, Université de Liège
Thomas Dubois, International Centre of Insect Physiology and Ecology Nairobi
Léna Durocher-Granger, CAB International (CABI)
José L. Fernández-Triana, Canadian National Collection of Insects, Arachnids and Nematodes
Frederic Francis, Université de Liège
Khalid Haddi, Universidade Federal de Lavras
Rhett D. Harrison, CIFOR-ICRAF
Muhammad Haseeb, Florida Agricultural and Mechanical University
Natasha S.A. Iwanicki, Universidade de São Paulo

Publication Date

2-15-2026

Abstract

Generative artificial intelligence (AI) could transform evidence synthesis and revolutionize the global scientific enterprise, yet its agricultural applications are understudied. Here, we systematically assess the performance of three web-grounded AI engines (ChatGPT, ScholarAI and DeepSeek) in synthesizing the global literature on biological control of the fall armyworm Spodoptera frugiperda, and benchmark their outputs against a recent, near-exhaustive human review. Though all engines rapidly screened vast literature corpora, they exhibited shortcomings in factual accuracy, reporting reliability and data consistency. In machine-run syntheses, natural enemy prevalence and performance data often diverged from published records while the level of agreement in enumerating top-performing taxa was evenly low. Meanwhile, internal consistency between laboratory and field-level parasitism data for ScholarAI and DeepSeek was similar to that in human-run reviews. All models tended towards faulty data extrapolation, hallucination and data fabrication, and a sporadic exclusion of key species. While autonomous, machine-only efforts accurately capture coarse-grained patterns in natural enemy identity, abundance, and impacts, they carry limited utility for (living) evidence syntheses or rigorous decision-support. Yet, handled with prudence and due human oversight, machine power might eventually revitalize underfunded disciplines and advance nature-friendly farming.

Keywords

Agroecology, Artificial intelligence, Biological control, Chatbot, Evidence synthesis, Large language models, Pesticide reduction, Systematic literature review

Recommended Citation

Wyckhuys, K., Akutse, K., Amalin, D., Araj, S., Beltran, M., Ben Fekih, I., Calatayud, P., Cicero, L., Cokola, M., Colmenarez, Y., Dessauvages, K., Dubois, T., Durocher-Granger, L., Fernández-Triana, J., Francis, F., Haddi, K., Harrison, R., Haseeb, M., & Iwanicki, N. (2026). Human versus machine: Can generative AI anticipate insect biological control outcomes?. Computers and Electronics in Agriculture, 242 https://doi.org/10.1016/j.compag.2025.111317